mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-24 10:24:35 +00:00
llama : fix BPE pre-tokenization (#6920)
* merged the changes from deepseeker models to main branch
* Moved regex patterns to unicode.cpp and updated unicode.h
* Moved header files
* Resolved issues
* added and refactored unicode_regex_split and related functions
* Updated/merged the deepseek coder pr
* Refactored code
* Adding unicode regex mappings
* Adding unicode regex function
* Added needed functionality, testing remains
* Fixed issues
* Fixed issue with gpt2 regex custom preprocessor
* unicode : fix? unicode_wstring_to_utf8
* lint : fix whitespaces
* tests : add tokenizer tests for numbers
* unicode : remove redundant headers
* tests : remove and rename tokenizer test scripts
* tests : add sample usage
* gguf-py : reader prints warnings on duplicate keys
* llama : towards llama3 tokenization support (wip)
* unicode : shot in the dark to fix tests on Windows
* unicode : first try custom implementations
* convert : add "tokenizer.ggml.pre" GGUF KV (wip)
* llama : use new pre-tokenizer type
* convert : fix pre-tokenizer type writing
* lint : fix
* make : add test-tokenizer-0-llama-v3
* wip
* models : add llama v3 vocab file
* llama : adapt punctuation regex + add llama 3 regex
* minor
* unicode : set bomb
* unicode : set bomb
* unicode : always use std::wregex
* unicode : support \p{N}, \p{L} and \p{P} natively
* unicode : try fix windows
* unicode : category support via std::regex
* unicode : clean-up
* unicode : simplify
* convert : add convert-hf-to-gguf-update.py
ggml-ci
* lint : update
* convert : add falcon
ggml-ci
* unicode : normalize signatures
* lint : fix
* lint : fix
* convert : remove unused functions
* convert : add comments
* convert : exercise contractions
ggml-ci
* lint : fix
* cmake : refactor test targets
* tests : refactor vocab tests
ggml-ci
* tests : add more vocabs and tests
ggml-ci
* unicode : cleanup
* scripts : ignore new update script in check-requirements.sh
* models : add phi-3, mpt, gpt-2, starcoder
* tests : disable obsolete
ggml-ci
* tests : use faster bpe test
ggml-ci
* llama : more prominent warning for old BPE models
* tests : disable test-tokenizer-1-bpe due to slowness
ggml-ci
---------
Co-authored-by: Jaggzh <jaggz.h@gmail.com>
Co-authored-by: Kazim Abrar Mahi <kazimabrarmahi135@gmail.com>
This commit is contained in:
Georgi Gerganov
GitHub
parent
3f167476b1
commit
f4ab2a4147
2
.github/workflows/python-lint.yml
vendored
2
.github/workflows/python-lint.yml
vendored
@ -21,4 +21,4 @@ jobs:
|
|||||||
uses: py-actions/flake8@v2
|
uses: py-actions/flake8@v2
|
||||||
with:
|
with:
|
||||||
ignore: "E203,E211,E221,E225,E231,E241,E251,E261,E266,E501,E701,E704,W503"
|
ignore: "E203,E211,E221,E225,E231,E241,E251,E261,E266,E501,E701,E704,W503"
|
||||||
exclude: "examples/*,examples/*/**,*/**/__init__.py"
|
exclude: "examples/*,examples/*/**,*/**/__init__.py,convert-hf-to-gguf-update.py"
|
||||||
|
|||||||
15
.gitignore
vendored
15
.gitignore
vendored
@ -108,3 +108,18 @@ examples/server/*.mjs.hpp
|
|||||||
poetry.lock
|
poetry.lock
|
||||||
poetry.toml
|
poetry.toml
|
||||||
nppBackup
|
nppBackup
|
||||||
|
|
||||||
|
# Test binaries
|
||||||
|
/tests/test-grammar-parser
|
||||||
|
/tests/test-llama-grammar
|
||||||
|
/tests/test-double-float
|
||||||
|
/tests/test-grad0
|
||||||
|
/tests/test-opt
|
||||||
|
/tests/test-quantize-fns
|
||||||
|
/tests/test-quantize-perf
|
||||||
|
/tests/test-sampling
|
||||||
|
/tests/test-tokenizer-0
|
||||||
|
/tests/test-tokenizer-1-spm
|
||||||
|
/tests/test-tokenizer-1-bpe
|
||||||
|
/tests/test-rope
|
||||||
|
/tests/test-backend-ops
|
||||||
|
|||||||
44
Makefile
44
Makefile
@ -6,11 +6,23 @@ BUILD_TARGETS = \
|
|||||||
|
|
||||||
# Binaries only useful for tests
|
# Binaries only useful for tests
|
||||||
TEST_TARGETS = \
|
TEST_TARGETS = \
|
||||||
tests/test-llama-grammar tests/test-grammar-parser tests/test-double-float tests/test-grad0 tests/test-opt \
|
tests/test-autorelease \
|
||||||
tests/test-quantize-fns tests/test-quantize-perf tests/test-sampling tests/test-tokenizer-0-llama \
|
tests/test-backend-ops \
|
||||||
tests/test-tokenizer-0-falcon tests/test-tokenizer-1-llama tests/test-tokenizer-1-bpe tests/test-rope \
|
tests/test-double-float \
|
||||||
tests/test-backend-ops tests/test-model-load-cancel tests/test-autorelease \
|
tests/test-grad0 \
|
||||||
tests/test-json-schema-to-grammar tests/test-grammar-integration
|
tests/test-grammar-integration \
|
||||||
|
tests/test-grammar-parser \
|
||||||
|
tests/test-json-schema-to-grammar \
|
||||||
|
tests/test-llama-grammar \
|
||||||
|
tests/test-model-load-cancel \
|
||||||
|
tests/test-opt \
|
||||||
|
tests/test-quantize-fns \
|
||||||
|
tests/test-quantize-perf \
|
||||||
|
tests/test-rope \
|
||||||
|
tests/test-sampling \
|
||||||
|
tests/test-tokenizer-0 \
|
||||||
|
tests/test-tokenizer-1-bpe \
|
||||||
|
tests/test-tokenizer-1-spm
|
||||||
|
|
||||||
# Code coverage output files
|
# Code coverage output files
|
||||||
COV_TARGETS = *.gcno tests/*.gcno *.gcda tests/*.gcda *.gcov tests/*.gcov lcov-report gcovr-report
|
COV_TARGETS = *.gcno tests/*.gcno *.gcda tests/*.gcda *.gcov tests/*.gcov lcov-report gcovr-report
|
||||||
@ -60,11 +72,17 @@ default: $(BUILD_TARGETS)
|
|||||||
test: $(TEST_TARGETS)
|
test: $(TEST_TARGETS)
|
||||||
@failures=0; \
|
@failures=0; \
|
||||||
for test_target in $(TEST_TARGETS); do \
|
for test_target in $(TEST_TARGETS); do \
|
||||||
if [ "$$test_target" = "tests/test-tokenizer-0-llama" ]; then \
|
if [ "$$test_target" = "tests/test-tokenizer-0" ]; then \
|
||||||
./$$test_target $(CURDIR)/models/ggml-vocab-llama.gguf; \
|
./$$test_target $(CURDIR)/models/ggml-vocab-llama-spm.gguf; \
|
||||||
elif [ "$$test_target" = "tests/test-tokenizer-0-falcon" ]; then \
|
./$$test_target $(CURDIR)/models/ggml-vocab-llama-bpe.gguf; \
|
||||||
|
./$$test_target $(CURDIR)/models/ggml-vocab-phi-3.gguf; \
|
||||||
./$$test_target $(CURDIR)/models/ggml-vocab-falcon.gguf; \
|
./$$test_target $(CURDIR)/models/ggml-vocab-falcon.gguf; \
|
||||||
elif [ "$$test_target" = "tests/test-tokenizer-1-llama" ]; then \
|
./$$test_target $(CURDIR)/models/ggml-vocab-deepseek-coder.gguf; \
|
||||||
|
./$$test_target $(CURDIR)/models/ggml-vocab-deepseek-llm.gguf; \
|
||||||
|
./$$test_target $(CURDIR)/models/ggml-vocab-bert-bge.gguf; \
|
||||||
|
./$$test_target $(CURDIR)/models/ggml-vocab-starcoder.gguf; \
|
||||||
|
./$$test_target $(CURDIR)/models/ggml-vocab-gpt-2.gguf; \
|
||||||
|
elif [ "$$test_target" = "tests/test-tokenizer-1-spm" ]; then \
|
||||||
continue; \
|
continue; \
|
||||||
elif [ "$$test_target" = "tests/test-tokenizer-1-bpe" ]; then \
|
elif [ "$$test_target" = "tests/test-tokenizer-1-bpe" ]; then \
|
||||||
continue; \
|
continue; \
|
||||||
@ -982,11 +1000,7 @@ tests/test-sampling: tests/test-sampling.cpp ggml.o llama.o $(OBJS)
|
|||||||
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
||||||
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
||||||
|
|
||||||
tests/test-tokenizer-0-falcon: tests/test-tokenizer-0-falcon.cpp ggml.o llama.o $(COMMON_DEPS) console.o $(OBJS)
|
tests/test-tokenizer-0: tests/test-tokenizer-0.cpp ggml.o llama.o $(COMMON_DEPS) console.o $(OBJS)
|
||||||
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
|
||||||
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
|
||||||
|
|
||||||
tests/test-tokenizer-0-llama: tests/test-tokenizer-0-llama.cpp ggml.o llama.o $(COMMON_DEPS) console.o $(OBJS)
|
|
||||||
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
||||||
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
||||||
|
|
||||||
@ -994,7 +1008,7 @@ tests/test-tokenizer-1-bpe: tests/test-tokenizer-1-bpe.cpp ggml.o llama.o $(COMM
|
|||||||
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
||||||
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
||||||
|
|
||||||
tests/test-tokenizer-1-llama: tests/test-tokenizer-1-llama.cpp ggml.o llama.o $(COMMON_DEPS) console.o $(OBJS)
|
tests/test-tokenizer-1-spm: tests/test-tokenizer-1-spm.cpp ggml.o llama.o $(COMMON_DEPS) console.o $(OBJS)
|
||||||
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
|
||||||
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
|
||||||
|
|
||||||
|
|||||||
@ -1693,6 +1693,18 @@ std::vector<std::string> string_split(std::string input, char separator) {
|
|||||||
return parts;
|
return parts;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
std::string string_strip(const std::string & str) {
|
||||||
|
size_t start = 0;
|
||||||
|
size_t end = str.size();
|
||||||
|
while (start < end && std::isspace(str[start])) {
|
||||||
|
start++;
|
||||||
|
}
|
||||||
|
while (end > start && std::isspace(str[end - 1])) {
|
||||||
|
end--;
|
||||||
|
}
|
||||||
|
return str.substr(start, end - start);
|
||||||
|
}
|
||||||
|
|
||||||
std::vector<llama_sampler_type> sampler_types_from_names(const std::vector<std::string> & names, bool allow_alt_names) {
|
std::vector<llama_sampler_type> sampler_types_from_names(const std::vector<std::string> & names, bool allow_alt_names) {
|
||||||
std::unordered_map<std::string, llama_sampler_type> sampler_canonical_name_map {
|
std::unordered_map<std::string, llama_sampler_type> sampler_canonical_name_map {
|
||||||
{"top_k", llama_sampler_type::TOP_K},
|
{"top_k", llama_sampler_type::TOP_K},
|
||||||
|
|||||||
@ -196,6 +196,7 @@ bool validate_file_name(const std::string & filename);
|
|||||||
std::vector<llama_sampler_type> sampler_types_from_names(const std::vector<std::string> & names, bool allow_alt_names);
|
std::vector<llama_sampler_type> sampler_types_from_names(const std::vector<std::string> & names, bool allow_alt_names);
|
||||||
std::vector<llama_sampler_type> sampler_types_from_chars(const std::string & names_string);
|
std::vector<llama_sampler_type> sampler_types_from_chars(const std::string & names_string);
|
||||||
std::vector<std::string> string_split(std::string input, char separator);
|
std::vector<std::string> string_split(std::string input, char separator);
|
||||||
|
std::string string_strip(const std::string & str);
|
||||||
std::string sampler_type_to_name_string(llama_sampler_type sampler_type);
|
std::string sampler_type_to_name_string(llama_sampler_type sampler_type);
|
||||||
|
|
||||||
//
|
//
|
||||||
|
|||||||
275
convert-hf-to-gguf-update.py
Normal file
275
convert-hf-to-gguf-update.py
Normal file
@ -0,0 +1,275 @@
|
|||||||
|
# This script downloads the tokenizer models of the specified models from Huggingface and
|
||||||
|
# generates the get_vocab_base_pre() function for convert-hf-to-gguf.py
|
||||||
|
#
|
||||||
|
# This is necessary in order to analyze the type of pre-tokenizer used by the model and
|
||||||
|
# provide the necessary information to llama.cpp via the GGUF header in order to implement
|
||||||
|
# the same pre-tokenizer.
|
||||||
|
#
|
||||||
|
# ref: https://github.com/ggerganov/llama.cpp/pull/6920
|
||||||
|
#
|
||||||
|
# Instructions:
|
||||||
|
#
|
||||||
|
# - Add a new model to the "models" list
|
||||||
|
# - Run the script with your huggingface token:
|
||||||
|
#
|
||||||
|
# python3 convert-hf-to-gguf-update.py <huggingface_token>
|
||||||
|
#
|
||||||
|
# - Copy-paste the generated get_vocab_base_pre() function into convert-hf-to-gguf.py
|
||||||
|
# - Update llama.cpp with the new pre-tokenizer if necessary
|
||||||
|
#
|
||||||
|
# TODO: generate tokenizer tests for llama.cpp
|
||||||
|
# TODO: automate the update of convert-hf-to-gguf.py
|
||||||
|
#
|
||||||
|
|
||||||
|
import os
|
||||||
|
import requests
|
||||||
|
import sys
|
||||||
|
import json
|
||||||
|
|
||||||
|
from hashlib import sha256
|
||||||
|
from enum import IntEnum, auto
|
||||||
|
|
||||||
|
class TOKENIZER_TYPE(IntEnum):
|
||||||
|
SPM = auto()
|
||||||
|
BPE = auto()
|
||||||
|
WPM = auto()
|
||||||
|
|
||||||
|
# TODO: this string has to exercise as much pre-tokenizer functionality as possible
|
||||||
|
# will be updated with time - contributions welcome
|
||||||
|
chktxt = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български \'\'\'\'\'\'```````\"\"\"\"......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL'
|
||||||
|
|
||||||
|
if len(sys.argv) == 2:
|
||||||
|
token = sys.argv[1]
|
||||||
|
else:
|
||||||
|
print("Usage: python convert-hf-to-gguf-update.py <huggingface_token>")
|
||||||
|
sys.exit(1)
|
||||||
|
|
||||||
|
# TODO: add models here, base models preferred
|
||||||
|
models = [
|
||||||
|
{ "name": "llama-spm", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/meta-llama/Llama-2-7b-hf", },
|
||||||
|
{ "name": "llama-bpe", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/meta-llama/Meta-Llama-3-8B", },
|
||||||
|
{ "name": "phi-3", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/microsoft/Phi-3-mini-4k-instruct", },
|
||||||
|
{ "name": "deepseek-llm", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/deepseek-llm-7b-base", },
|
||||||
|
{ "name": "deepseek-coder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base", },
|
||||||
|
{ "name": "falcon", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/tiiuae/falcon-7b", },
|
||||||
|
{ "name": "bert-bge", "tokt": TOKENIZER_TYPE.WPM, "repo": "https://huggingface.co/BAAI/bge-small-en-v1.5", },
|
||||||
|
{ "name": "mpt", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/mosaicml/mpt-7b", },
|
||||||
|
{ "name": "starcoder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/bigcode/starcoder2-3b", },
|
||||||
|
{ "name": "gpt-2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/openai-community/gpt2", },
|
||||||
|
]
|
||||||
|
|
||||||
|
# make directory "models/tokenizers" if it doesn't exist
|
||||||
|
if not os.path.exists("models/tokenizers"):
|
||||||
|
os.makedirs("models/tokenizers")
|
||||||
|
|
||||||
|
def download_file_with_auth(url, token, save_path):
|
||||||
|
headers = {"Authorization": f"Bearer {token}"}
|
||||||
|
response = requests.get(url, headers=headers)
|
||||||
|
if response.status_code == 200:
|
||||||
|
with open(save_path, 'wb') as f:
|
||||||
|
f.write(response.content)
|
||||||
|
print(f"File {save_path} downloaded successfully")
|
||||||
|
else:
|
||||||
|
print(f"Failed to download file. Status code: {response.status_code}")
|
||||||
|
|
||||||
|
# download the tokenizer models
|
||||||
|
for model in models:
|
||||||
|
name = model["name"]
|
||||||
|
repo = model["repo"]
|
||||||
|
tokt = model["tokt"]
|
||||||
|
|
||||||
|
if not os.path.exists(f"models/tokenizers/{name}"):
|
||||||
|
os.makedirs(f"models/tokenizers/{name}")
|
||||||
|
else:
|
||||||
|
print(f"Directory models/tokenizers/{name} already exists - skipping")
|
||||||
|
continue
|
||||||
|
|
||||||
|
print(f"Downloading {name} to models/tokenizers/{name}")
|
||||||
|
|
||||||
|
url = f"{repo}/raw/main/config.json"
|
||||||
|
save_path = f"models/tokenizers/{name}/config.json"
|
||||||
|
download_file_with_auth(url, token, save_path)
|
||||||
|
|
||||||
|
url = f"{repo}/raw/main/tokenizer.json"
|
||||||
|
save_path = f"models/tokenizers/{name}/tokenizer.json"
|
||||||
|
download_file_with_auth(url, token, save_path)
|
||||||
|
|
||||||
|
if tokt == TOKENIZER_TYPE.SPM:
|
||||||
|
url = f"{repo}/resolve/main/tokenizer.model"
|
||||||
|
save_path = f"models/tokenizers/{name}/tokenizer.model"
|
||||||
|
download_file_with_auth(url, token, save_path)
|
||||||
|
|
||||||
|
url = f"{repo}/raw/main/tokenizer_config.json"
|
||||||
|
save_path = f"models/tokenizers/{name}/tokenizer_config.json"
|
||||||
|
download_file_with_auth(url, token, save_path)
|
||||||
|
|
||||||
|
# generate the source code for the convert-hf-to-gguf.py:get_vocab_base_pre() function:
|
||||||
|
# TODO: auto-update convert-hf-to-gguf.py with the generated function
|
||||||
|
|
||||||
|
src_ifs = ""
|
||||||
|
for model in models:
|
||||||
|
name = model["name"]
|
||||||
|
tokt = model["tokt"]
|
||||||
|
|
||||||
|
if tokt == TOKENIZER_TYPE.SPM:
|
||||||
|
continue
|
||||||
|
|
||||||
|
# create the tokenizer
|
||||||
|
from transformers import AutoTokenizer
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}")
|
||||||
|
|
||||||
|
chktok = tokenizer.encode(chktxt)
|
||||||
|
chkhsh = sha256(str(chktok).encode()).hexdigest()
|
||||||
|
|
||||||
|
print(f"model: {name}")
|
||||||
|
print(f"tokt: {tokt}")
|
||||||
|
print(f"repo: {model['repo']}")
|
||||||
|
print(f"chktok: {chktok}")
|
||||||
|
print(f"chkhsh: {chkhsh}")
|
||||||
|
|
||||||
|
# print the "pre_tokenizer" content from the tokenizer.json
|
||||||
|
with open(f"models/tokenizers/{name}/tokenizer.json", "r") as f:
|
||||||
|
cfg = json.load(f)
|
||||||
|
pre_tokenizer = cfg["pre_tokenizer"]
|
||||||
|
print("pre_tokenizer: " + json.dumps(pre_tokenizer, indent=4))
|
||||||
|
|
||||||
|
print(f"\n")
|
||||||
|
|
||||||
|
src_ifs += f" if chkhsh == \"{chkhsh}\":\n"
|
||||||
|
src_ifs += f" # ref: {model['repo']}\n"
|
||||||
|
src_ifs += f" res = \"{name}\"\n"
|
||||||
|
|
||||||
|
src_func = ""
|
||||||
|
src_func += " def get_vocab_base_pre(self, tokenizer) -> str:\n"
|
||||||
|
src_func += " # encoding this string and hashing the resulting tokens would (hopefully) give us a unique identifier that\n"
|
||||||
|
src_func += " # is specific for the BPE pre-tokenizer used by the model\n"
|
||||||
|
src_func += " # we will use this unique identifier to write a \"tokenizer.ggml.pre\" entry in the GGUF file which we can\n"
|
||||||
|
src_func += " # use in llama.cpp to implement the same pre-tokenizer\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += f" chktxt = {repr(chktxt)}\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " chktok = tokenizer.encode(chktxt)\n"
|
||||||
|
src_func += " chkhsh = sha256(str(chktok).encode()).hexdigest()\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " print(f\"chktok: {chktok}\")\n"
|
||||||
|
src_func += " print(f\"chkhsh: {chkhsh}\")\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " res = None\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " # NOTE: if you get an error here, you need to add the model to the if-elif chain below\n"
|
||||||
|
src_func += " # don't do this manually - use the convert-hf-to-gguf-update.py script!\n"
|
||||||
|
src_func += f"{src_ifs}\n"
|
||||||
|
src_func += " if res is None:\n"
|
||||||
|
src_func += " print(\"\\n\")\n"
|
||||||
|
src_func += " print(\"**************************************************************************************\")\n"
|
||||||
|
src_func += " print(\"** WARNING: The BPE pre-tokenizer was not recognized!\")\n"
|
||||||
|
src_func += " print(\"** This means that it was not added yet or you are using an older version.\")\n"
|
||||||
|
src_func += " print(\"** Check convert-hf-to-gguf-update.py and update it accordingly.\")\n"
|
||||||
|
src_func += " print(\"**\")\n"
|
||||||
|
src_func += " print(f\"** chkhsh: {chkhsh}\")\n"
|
||||||
|
src_func += " print(\"**************************************************************************************\")\n"
|
||||||
|
src_func += " print(\"\\n\")\n"
|
||||||
|
src_func += " raise NotImplementedError(\"BPE pre-tokenizer was not recognized - update get_vocab_base_pre()\")\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " print(f\"tokenizer.ggml.pre: {res}\")\n"
|
||||||
|
src_func += " print(f\"chkhsh: {chkhsh}\")\n"
|
||||||
|
src_func += "\n"
|
||||||
|
src_func += " return res\n"
|
||||||
|
|
||||||
|
print(src_func)
|
||||||
|
|
||||||
|
print("\n")
|
||||||
|
print("!!! Copy-paste the function above into convert-hf-to-gguf.py !!!")
|
||||||
|
print("\n")
|
||||||
|
|
||||||
|
# generate tests for each tokenizer model
|
||||||
|
|
||||||
|
tests = [
|
||||||
|
"",
|
||||||
|
" ",
|
||||||
|
" ",
|
||||||
|
" ",
|
||||||
|
"\t",
|
||||||
|
"\n",
|
||||||
|
"\n\n",
|
||||||
|
"\n\n\n",
|
||||||
|
"\t\n",
|
||||||
|
"Hello world",
|
||||||
|
" Hello world",
|
||||||
|
"Hello World",
|
||||||
|
" Hello World",
|
||||||
|
" Hello World!",
|
||||||
|
"Hello, world!",
|
||||||
|
" Hello, world!",
|
||||||
|
" this is 🦙.cpp",
|
||||||
|
"w048 7tuijk dsdfhu",
|
||||||
|
"нещо на Български",
|
||||||
|
"កាន់តែពិសេសអាចខលចេញ",
|
||||||
|
"🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)",
|
||||||
|
"Hello",
|
||||||
|
" Hello",
|
||||||
|
" Hello",
|
||||||
|
" Hello",
|
||||||
|
" Hello",
|
||||||
|
" Hello\n Hello",
|
||||||
|
" (",
|
||||||
|
"\n =",
|
||||||
|
"' era",
|
||||||
|
"Hello, y'all! How are you 😁 ?我想在apple工作1314151天～",
|
||||||
|
"3",
|
||||||
|
"33",
|
||||||
|
"333",
|
||||||
|
"3333",
|
||||||
|
"33333",
|
||||||
|
"333333",
|
||||||
|
"3333333",
|
||||||
|
"33333333",
|
||||||
|
"333333333",
|
||||||
|
chktxt,
|
||||||
|
]
|
||||||
|
|
||||||
|
# write the tests to ./models/ggml-vocab-{name}.gguf.inp
|
||||||
|
# the format is:
|
||||||
|
#
|
||||||
|
# test0
|
||||||
|
# __ggml_vocab_test__
|
||||||
|
# test1
|
||||||
|
# __ggml_vocab_test__
|
||||||
|
# ...
|
||||||
|
#
|
||||||
|
|
||||||
|
# with each model, encode all tests and write the results in ./models/ggml-vocab-{name}.gguf.out
|
||||||
|
# for each test, write the resulting tokens on a separate line
|
||||||
|
|
||||||
|
for model in models:
|
||||||
|
name = model["name"]
|
||||||
|
tokt = model["tokt"]
|
||||||
|
|
||||||
|
# create the tokenizer
|
||||||
|
from transformers import AutoTokenizer
|
||||||
|
tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}")
|
||||||
|
|
||||||
|
with open(f"models/ggml-vocab-{name}.gguf.inp", "w") as f:
|
||||||
|
for text in tests:
|
||||||
|
f.write(f"{text}")
|
||||||
|
f.write("\n__ggml_vocab_test__\n")
|
||||||
|
|
||||||
|
with open(f"models/ggml-vocab-{name}.gguf.out", "w") as f:
|
||||||
|
for text in tests:
|
||||||
|
res = tokenizer.encode(text, add_special_tokens=False)
|
||||||
|
for r in res:
|
||||||
|
f.write(f" {r}")
|
||||||
|
f.write("\n")
|
||||||
|
|
||||||
|
print(f"Tests for {name} written in ./models/ggml-vocab-{name}.gguf.*")
|
||||||
|
|
||||||
|
# generate commands for creating vocab files
|
||||||
|
|
||||||
|
print("\nRun the following commands to generate the vocab files for testing:\n")
|
||||||
|
|
||||||
|
for model in models:
|
||||||
|
name = model["name"]
|
||||||
|
|
||||||
|
print(f"python3 convert-hf-to-gguf.py models/tokenizers/{name}/ --outfile models/ggml-vocab-{name}.gguf --vocab-only")
|
||||||
|
|
||||||
|
print("\n")
|
||||||
@ -11,6 +11,7 @@ import sys
|
|||||||
from abc import ABC, abstractmethod
|
from abc import ABC, abstractmethod
|
||||||
from enum import IntEnum
|
from enum import IntEnum
|
||||||
from pathlib import Path
|
from pathlib import Path
|
||||||
|
from hashlib import sha256
|
||||||
from typing import TYPE_CHECKING, Any, Callable, ContextManager, Iterator, Sequence, TypeVar, cast
|
from typing import TYPE_CHECKING, Any, Callable, ContextManager, Iterator, Sequence, TypeVar, cast
|
||||||
|
|
||||||
import numpy as np
|
import numpy as np
|
||||||
@ -229,7 +230,7 @@ class Model(ABC):
|
|||||||
return (f"pytorch_model-{n:05}-of-{self.num_parts:05}.bin" for n in range(1, self.num_parts + 1))
|
return (f"pytorch_model-{n:05}-of-{self.num_parts:05}.bin" for n in range(1, self.num_parts + 1))
|
||||||
|
|
||||||
# used for GPT-2 BPE and WordPiece vocabs
|
# used for GPT-2 BPE and WordPiece vocabs
|
||||||
def get_basic_vocab(self) -> tuple[list[str], list[int]]:
|
def get_vocab_base(self) -> tuple[list[str], list[int], str]:
|
||||||
tokens: list[str] = []
|
tokens: list[str] = []
|
||||||
toktypes: list[int] = []
|
toktypes: list[int] = []
|
||||||
|
|
||||||
@ -238,6 +239,8 @@ class Model(ABC):
|
|||||||
vocab_size = self.hparams.get("vocab_size", len(tokenizer.vocab))
|
vocab_size = self.hparams.get("vocab_size", len(tokenizer.vocab))
|
||||||
assert max(tokenizer.vocab.values()) < vocab_size
|
assert max(tokenizer.vocab.values()) < vocab_size
|
||||||
|
|
||||||
|
tokpre = self.get_vocab_base_pre(tokenizer)
|
||||||
|
|
||||||
reverse_vocab = {id_: encoded_tok for encoded_tok, id_ in tokenizer.vocab.items()}
|
reverse_vocab = {id_: encoded_tok for encoded_tok, id_ in tokenizer.vocab.items()}
|
||||||
added_vocab = tokenizer.get_added_vocab()
|
added_vocab = tokenizer.get_added_vocab()
|
||||||
|
|
||||||
@ -255,11 +258,75 @@ class Model(ABC):
|
|||||||
tokens.append(reverse_vocab[i])
|
tokens.append(reverse_vocab[i])
|
||||||
toktypes.append(gguf.TokenType.NORMAL)
|
toktypes.append(gguf.TokenType.NORMAL)
|
||||||
|
|
||||||
return tokens, toktypes
|
return tokens, toktypes, tokpre
|
||||||
|
|
||||||
|
# NOTE: this function is generated by convert-hf-to-gguf-update.py
|
||||||
|
# do not modify it manually!
|
||||||
|
# ref: https://github.com/ggerganov/llama.cpp/pull/6920
|
||||||
|
def get_vocab_base_pre(self, tokenizer) -> str:
|
||||||
|
# encoding this string and hashing the resulting tokens would (hopefully) give us a unique identifier that
|
||||||
|
# is specific for the BPE pre-tokenizer used by the model
|
||||||
|
# we will use this unique identifier to write a "tokenizer.ggml.pre" entry in the GGUF file which we can
|
||||||
|
# use in llama.cpp to implement the same pre-tokenizer
|
||||||
|
|
||||||
|
chktxt = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶\u200d🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български \'\'\'\'\'\'```````""""......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL'
|
||||||
|
|
||||||
|
chktok = tokenizer.encode(chktxt)
|
||||||
|
chkhsh = sha256(str(chktok).encode()).hexdigest()
|
||||||
|
|
||||||
|
print(f"chktok: {chktok}")
|
||||||
|
print(f"chkhsh: {chkhsh}")
|
||||||
|
|
||||||
|
res = None
|
||||||
|
|
||||||
|
# NOTE: if you get an error here, you need to add the model to the if-elif chain below
|
||||||
|
# don't do this manually - use the convert-hf-to-gguf-update.py script!
|
||||||
|
if chkhsh == "0ef9807a4087ebef797fc749390439009c3b9eda9ad1a097abbe738f486c01e5":
|
||||||
|
# ref: https://huggingface.co/meta-llama/Meta-Llama-3-8B
|
||||||
|
res = "llama-bpe"
|
||||||
|
if chkhsh == "049ecf7629871e3041641907f3de7c733e4dbfdc736f57d882ba0b0845599754":
|
||||||
|
# ref: https://huggingface.co/deepseek-ai/deepseek-llm-7b-base
|
||||||
|
res = "deepseek-llm"
|
||||||
|
if chkhsh == "347715f544604f9118bb75ed199f68779f423cabb20db6de6f31b908d04d7821":
|
||||||
|
# ref: https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base
|
||||||
|
res = "deepseek-coder"
|
||||||
|
if chkhsh == "8aeee3860c56296a157a1fe2fad249ec40aa59b1bb5709f4ade11c4e6fe652ed":
|
||||||
|
# ref: https://huggingface.co/tiiuae/falcon-7b
|
||||||
|
res = "falcon"
|
||||||
|
if chkhsh == "0876d13b50744004aa9aeae05e7b0647eac9d801b5ba4668afc01e709c15e19f":
|
||||||
|
# ref: https://huggingface.co/BAAI/bge-small-en-v1.5
|
||||||
|
res = "bert-bge"
|
||||||
|
if chkhsh == "b6dc8df998e1cfbdc4eac8243701a65afe638679230920b50d6f17d81c098166":
|
||||||
|
# ref: https://huggingface.co/mosaicml/mpt-7b
|
||||||
|
res = "mpt"
|
||||||
|
if chkhsh == "35d91631860c815f952d711435f48d356ebac988362536bed955d43bfa436e34":
|
||||||
|
# ref: https://huggingface.co/bigcode/starcoder2-3b
|
||||||
|
res = "starcoder"
|
||||||
|
if chkhsh == "3ce83efda5659b07b1ad37ca97ca5797ea4285d9b9ab0dc679e4a720c9da7454":
|
||||||
|
# ref: https://huggingface.co/openai-community/gpt2
|
||||||
|
res = "gpt-2"
|
||||||
|
|
||||||
|
if res is None:
|
||||||
|
print("\n")
|
||||||
|
print("**************************************************************************************")
|
||||||
|
print("** WARNING: The BPE pre-tokenizer was not recognized!")
|
||||||
|
print("** This means that it was not added yet or you are using an older version.")
|
||||||
|
print("** Check convert-hf-to-gguf-update.py and update it accordingly.")
|
||||||
|
print("**")
|
||||||
|
print(f"** chkhsh: {chkhsh}")
|
||||||
|
print("**************************************************************************************")
|
||||||
|
print("\n")
|
||||||
|
raise NotImplementedError("BPE pre-tokenizer was not recognized - update get_vocab_base_pre()")
|
||||||
|
|
||||||
|
print(f"tokenizer.ggml.pre: {res}")
|
||||||
|
print(f"chkhsh: {chkhsh}")
|
||||||
|
|
||||||
|
return res
|
||||||
|
|
||||||
def _set_vocab_gpt2(self) -> None:
|
def _set_vocab_gpt2(self) -> None:
|
||||||
tokens, toktypes = self.get_basic_vocab()
|
tokens, toktypes, tokpre = self.get_vocab_base()
|
||||||
self.gguf_writer.add_tokenizer_model("gpt2")
|
self.gguf_writer.add_tokenizer_model("gpt2")
|
||||||
|
self.gguf_writer.add_tokenizer_pre(tokpre)
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
|
|
||||||
@ -277,6 +344,8 @@ class Model(ABC):
|
|||||||
vocab_size = hparams["vocab_size"]
|
vocab_size = hparams["vocab_size"]
|
||||||
assert max(tokenizer.get_vocab().values()) < vocab_size
|
assert max(tokenizer.get_vocab().values()) < vocab_size
|
||||||
|
|
||||||
|
tokpre = self.get_vocab_base_pre(tokenizer)
|
||||||
|
|
||||||
merges = []
|
merges = []
|
||||||
vocab = {}
|
vocab = {}
|
||||||
mergeable_ranks = tokenizer.mergeable_ranks
|
mergeable_ranks = tokenizer.mergeable_ranks
|
||||||
@ -304,6 +373,7 @@ class Model(ABC):
|
|||||||
toktypes.append(gguf.TokenType.NORMAL)
|
toktypes.append(gguf.TokenType.NORMAL)
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("gpt2")
|
self.gguf_writer.add_tokenizer_model("gpt2")
|
||||||
|
self.gguf_writer.add_tokenizer_pre(tokpre)
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
|
|
||||||
@ -376,6 +446,7 @@ class Model(ABC):
|
|||||||
assert len(tokens) == vocab_size
|
assert len(tokens) == vocab_size
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("llama")
|
self.gguf_writer.add_tokenizer_model("llama")
|
||||||
|
self.gguf_writer.add_tokenizer_pre("default")
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_scores(scores)
|
self.gguf_writer.add_token_scores(scores)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
@ -397,6 +468,7 @@ class Model(ABC):
|
|||||||
assert len(tokens) == vocab.vocab_size
|
assert len(tokens) == vocab.vocab_size
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("llama")
|
self.gguf_writer.add_tokenizer_model("llama")
|
||||||
|
self.gguf_writer.add_tokenizer_pre("default")
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_scores(scores)
|
self.gguf_writer.add_token_scores(scores)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
@ -840,6 +912,7 @@ class XverseModel(Model):
|
|||||||
toktypes.append(toktype)
|
toktypes.append(toktype)
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("llama")
|
self.gguf_writer.add_tokenizer_model("llama")
|
||||||
|
self.gguf_writer.add_tokenizer_pre("default")
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
|
|
||||||
@ -1335,6 +1408,11 @@ class LlamaModel(Model):
|
|||||||
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
|
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
|
||||||
self.gguf_writer.add_rope_dimension_count(hparams["hidden_size"] // hparams["num_attention_heads"])
|
self.gguf_writer.add_rope_dimension_count(hparams["hidden_size"] // hparams["num_attention_heads"])
|
||||||
|
|
||||||
|
if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
|
||||||
|
if self.hparams["rope_scaling"].get("type") == "linear":
|
||||||
|
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.LINEAR)
|
||||||
|
self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
|
||||||
|
|
||||||
# Same as super class, but permuting q_proj, k_proj
|
# Same as super class, but permuting q_proj, k_proj
|
||||||
def write_tensors(self):
|
def write_tensors(self):
|
||||||
block_count = self.hparams.get("n_layers", self.hparams.get("num_hidden_layers", self.hparams.get("n_layer")))
|
block_count = self.hparams.get("n_layers", self.hparams.get("num_hidden_layers", self.hparams.get("n_layer")))
|
||||||
@ -2052,6 +2130,7 @@ class Phi3MiniModel(Model):
|
|||||||
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
|
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("llama")
|
self.gguf_writer.add_tokenizer_model("llama")
|
||||||
|
self.gguf_writer.add_tokenizer_pre("default")
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_scores(scores)
|
self.gguf_writer.add_token_scores(scores)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
@ -2294,6 +2373,7 @@ class InternLM2Model(Model):
|
|||||||
toktypes.append(SentencePieceTokenTypes.USER_DEFINED)
|
toktypes.append(SentencePieceTokenTypes.USER_DEFINED)
|
||||||
|
|
||||||
self.gguf_writer.add_tokenizer_model("llama")
|
self.gguf_writer.add_tokenizer_model("llama")
|
||||||
|
self.gguf_writer.add_tokenizer_pre("default")
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_scores(scores)
|
self.gguf_writer.add_token_scores(scores)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
@ -2443,7 +2523,7 @@ class BertModel(Model):
|
|||||||
self.gguf_writer.add_pooling_type(pooling_type)
|
self.gguf_writer.add_pooling_type(pooling_type)
|
||||||
|
|
||||||
def set_vocab(self):
|
def set_vocab(self):
|
||||||
tokens, toktypes = self.get_basic_vocab()
|
tokens, toktypes, tokpre = self.get_vocab_base()
|
||||||
self.vocab_size = len(tokens)
|
self.vocab_size = len(tokens)
|
||||||
|
|
||||||
# we need this to validate the size of the token_type embeddings
|
# we need this to validate the size of the token_type embeddings
|
||||||
@ -2461,6 +2541,7 @@ class BertModel(Model):
|
|||||||
|
|
||||||
# add vocab to gguf
|
# add vocab to gguf
|
||||||
self.gguf_writer.add_tokenizer_model("bert")
|
self.gguf_writer.add_tokenizer_model("bert")
|
||||||
|
self.gguf_writer.add_tokenizer_pre(tokpre)
|
||||||
self.gguf_writer.add_token_list(tokens)
|
self.gguf_writer.add_token_list(tokens)
|
||||||
self.gguf_writer.add_token_types(toktypes)
|
self.gguf_writer.add_token_types(toktypes)
|
||||||
|
|
||||||
@ -2642,6 +2723,9 @@ class MambaModel(Model):
|
|||||||
field = neox_reader.get_field(gguf.Keys.Tokenizer.MODEL)
|
field = neox_reader.get_field(gguf.Keys.Tokenizer.MODEL)
|
||||||
self.gguf_writer.add_tokenizer_model(bytes(field.parts[-1]))
|
self.gguf_writer.add_tokenizer_model(bytes(field.parts[-1]))
|
||||||
|
|
||||||
|
field = neox_reader.get_field(gguf.Keys.Tokenizer.PRE)
|
||||||
|
self.gguf_writer.add_tokenizer_pre(bytes(field.parts[-1]))
|
||||||
|
|
||||||
field = neox_reader.get_field(gguf.Keys.Tokenizer.LIST)
|
field = neox_reader.get_field(gguf.Keys.Tokenizer.LIST)
|
||||||
self.gguf_writer.add_token_list([bytes(field.parts[i]) for i in field.data][:vocab_size])
|
self.gguf_writer.add_token_list([bytes(field.parts[i]) for i in field.data][:vocab_size])
|
||||||
|
|
||||||
@ -2847,6 +2931,7 @@ def parse_args() -> argparse.Namespace:
|
|||||||
help="directory containing model file",
|
help="directory containing model file",
|
||||||
)
|
)
|
||||||
parser.add_argument("--use-temp-file", action="store_true", help="use the tempfile library while processing (helpful when running out of memory, process killed)")
|
parser.add_argument("--use-temp-file", action="store_true", help="use the tempfile library while processing (helpful when running out of memory, process killed)")
|
||||||
|
parser.add_argument("--model-name", type=str, default=None, help="name of the model")
|
||||||
|
|
||||||
return parser.parse_args()
|
return parser.parse_args()
|
||||||
|
|
||||||
|
|||||||
@ -281,6 +281,7 @@ class GGMLToGGUF:
|
|||||||
def add_vocab(self, gguf_writer):
|
def add_vocab(self, gguf_writer):
|
||||||
hp = self.model.hyperparameters
|
hp = self.model.hyperparameters
|
||||||
gguf_writer.add_tokenizer_model('llama')
|
gguf_writer.add_tokenizer_model('llama')
|
||||||
|
gguf_writer.add_tokenizer_pre('default')
|
||||||
tokens = []
|
tokens = []
|
||||||
scores = []
|
scores = []
|
||||||
toktypes = []
|
toktypes = []
|
||||||
|
|||||||
@ -99,6 +99,7 @@ def main():
|
|||||||
|
|
||||||
tokens, scores, toktypes = _get_sentencepiece_tokenizer_info(args.model_dir)
|
tokens, scores, toktypes = _get_sentencepiece_tokenizer_info(args.model_dir)
|
||||||
gguf_writer.add_tokenizer_model('llama')
|
gguf_writer.add_tokenizer_model('llama')
|
||||||
|
gguf_writer.add_tokenizer_pre('default')
|
||||||
gguf_writer.add_token_list(tokens)
|
gguf_writer.add_token_list(tokens)
|
||||||
gguf_writer.add_token_scores(scores)
|
gguf_writer.add_token_scores(scores)
|
||||||
gguf_writer.add_token_types(toktypes)
|
gguf_writer.add_token_types(toktypes)
|
||||||
|
|||||||
@ -72,6 +72,7 @@ class Keys:
|
|||||||
|
|
||||||
class Tokenizer:
|
class Tokenizer:
|
||||||
MODEL = "tokenizer.ggml.model"
|
MODEL = "tokenizer.ggml.model"
|
||||||
|
PRE = "tokenizer.ggml.pre"
|
||||||
LIST = "tokenizer.ggml.tokens"
|
LIST = "tokenizer.ggml.tokens"
|
||||||
TOKEN_TYPE = "tokenizer.ggml.token_type"
|
TOKEN_TYPE = "tokenizer.ggml.token_type"
|
||||||
TOKEN_TYPE_COUNT = "tokenizer.ggml.token_type_count" # for BERT-style token types
|
TOKEN_TYPE_COUNT = "tokenizer.ggml.token_type_count" # for BERT-style token types
|
||||||
@ -940,6 +941,7 @@ KEY_SSM_TIME_STEP_RANK = Keys.SSM.TIME_STEP_RANK
|
|||||||
|
|
||||||
# tokenization
|
# tokenization
|
||||||
KEY_TOKENIZER_MODEL = Keys.Tokenizer.MODEL
|
KEY_TOKENIZER_MODEL = Keys.Tokenizer.MODEL
|
||||||
|
KEY_TOKENIZER_PRE = Keys.Tokenizer.PRE
|
||||||
KEY_TOKENIZER_LIST = Keys.Tokenizer.LIST
|
KEY_TOKENIZER_LIST = Keys.Tokenizer.LIST
|
||||||
KEY_TOKENIZER_TOKEN_TYPE = Keys.Tokenizer.TOKEN_TYPE
|
KEY_TOKENIZER_TOKEN_TYPE = Keys.Tokenizer.TOKEN_TYPE
|
||||||
KEY_TOKENIZER_SCORES = Keys.Tokenizer.SCORES
|
KEY_TOKENIZER_SCORES = Keys.Tokenizer.SCORES
|
||||||
|
|||||||
@ -139,8 +139,13 @@ class GGUFReader:
|
|||||||
|
|
||||||
def _push_field(self, field: ReaderField, skip_sum: bool = False) -> int:
|
def _push_field(self, field: ReaderField, skip_sum: bool = False) -> int:
|
||||||
if field.name in self.fields:
|
if field.name in self.fields:
|
||||||
raise KeyError(f'Duplicate {field.name} already in list at offset {field.offset}')
|
# TODO: add option to generate error on duplicate keys
|
||||||
self.fields[field.name] = field
|
# raise KeyError(f'Duplicate {field.name} already in list at offset {field.offset}')
|
||||||
|
|
||||||
|
print(f'Warning: Duplicate key {field.name} at offset {field.offset}')
|
||||||
|
self.fields[field.name + '_{}'.format(field.offset)] = field
|
||||||
|
else:
|
||||||
|
self.fields[field.name] = field
|
||||||
return 0 if skip_sum else sum(int(part.nbytes) for part in field.parts)
|
return 0 if skip_sum else sum(int(part.nbytes) for part in field.parts)
|
||||||
|
|
||||||
def _get_str(self, offset: int) -> tuple[npt.NDArray[np.uint64], npt.NDArray[np.uint8]]:
|
def _get_str(self, offset: int) -> tuple[npt.NDArray[np.uint64], npt.NDArray[np.uint8]]:
|
||||||
|
|||||||
@ -427,6 +427,9 @@ class GGUFWriter:
|
|||||||
def add_tokenizer_model(self, model: str) -> None:
|
def add_tokenizer_model(self, model: str) -> None:
|
||||||
self.add_string(Keys.Tokenizer.MODEL, model)
|
self.add_string(Keys.Tokenizer.MODEL, model)
|
||||||
|
|
||||||
|
def add_tokenizer_pre(self, pre: str) -> None:
|
||||||
|
self.add_string(Keys.Tokenizer.PRE, pre)
|
||||||
|
|
||||||
def add_token_list(self, tokens: Sequence[str] | Sequence[bytes] | Sequence[bytearray]) -> None:
|
def add_token_list(self, tokens: Sequence[str] | Sequence[bytes] | Sequence[bytearray]) -> None:
|
||||||
self.add_array(Keys.Tokenizer.LIST, tokens)
|
self.add_array(Keys.Tokenizer.LIST, tokens)
|
||||||
|
|
||||||
|
|||||||
311
llama.cpp
311
llama.cpp
@ -317,6 +317,7 @@ enum llm_kv {
|
|||||||
LLM_KV_SSM_TIME_STEP_RANK,
|
LLM_KV_SSM_TIME_STEP_RANK,
|
||||||
|
|
||||||
LLM_KV_TOKENIZER_MODEL,
|
LLM_KV_TOKENIZER_MODEL,
|
||||||
|
LLM_KV_TOKENIZER_PRE,
|
||||||
LLM_KV_TOKENIZER_LIST,
|
LLM_KV_TOKENIZER_LIST,
|
||||||
LLM_KV_TOKENIZER_TOKEN_TYPE,
|
LLM_KV_TOKENIZER_TOKEN_TYPE,
|
||||||
LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT,
|
LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT,
|
||||||
@ -393,6 +394,7 @@ static const std::map<llm_kv, const char *> LLM_KV_NAMES = {
|
|||||||
{ LLM_KV_SSM_TIME_STEP_RANK, "%s.ssm.time_step_rank" },
|
{ LLM_KV_SSM_TIME_STEP_RANK, "%s.ssm.time_step_rank" },
|
||||||
|
|
||||||
{ LLM_KV_TOKENIZER_MODEL, "tokenizer.ggml.model" },
|
{ LLM_KV_TOKENIZER_MODEL, "tokenizer.ggml.model" },
|
||||||
|
{ LLM_KV_TOKENIZER_PRE, "tokenizer.ggml.pre" },
|
||||||
{ LLM_KV_TOKENIZER_LIST, "tokenizer.ggml.tokens" },
|
{ LLM_KV_TOKENIZER_LIST, "tokenizer.ggml.tokens" },
|
||||||
{ LLM_KV_TOKENIZER_TOKEN_TYPE, "tokenizer.ggml.token_type" },
|
{ LLM_KV_TOKENIZER_TOKEN_TYPE, "tokenizer.ggml.token_type" },
|
||||||
{ LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, "tokenizer.ggml.token_type_count" },
|
{ LLM_KV_TOKENIZER_TOKEN_TYPE_COUNT, "tokenizer.ggml.token_type_count" },
|
||||||
@ -2115,7 +2117,8 @@ struct llama_vocab {
|
|||||||
ttype type;
|
ttype type;
|
||||||
};
|
};
|
||||||
|
|
||||||
enum llama_vocab_type type = LLAMA_VOCAB_TYPE_SPM;
|
enum llama_vocab_type type = LLAMA_VOCAB_TYPE_SPM;
|
||||||
|
enum llama_vocab_pre_type type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
||||||
|
|
||||||
std::unordered_map<token, id> token_to_id;
|
std::unordered_map<token, id> token_to_id;
|
||||||
std::vector<token_data> id_to_token;
|
std::vector<token_data> id_to_token;
|
||||||
@ -4214,11 +4217,13 @@ static void llm_load_vocab(
|
|||||||
|
|
||||||
// determine vocab type
|
// determine vocab type
|
||||||
{
|
{
|
||||||
std::string tokenizer_name;
|
std::string tokenizer_model;
|
||||||
|
std::string tokenizer_pre;
|
||||||
|
|
||||||
ml.get_key(LLM_KV_TOKENIZER_MODEL, tokenizer_name);
|
ml.get_key(LLM_KV_TOKENIZER_MODEL, tokenizer_model);
|
||||||
|
ml.get_key(LLM_KV_TOKENIZER_PRE, tokenizer_pre, false);
|
||||||
|
|
||||||
if (tokenizer_name == "no_vocab") {
|
if (tokenizer_model == "no_vocab") {
|
||||||
vocab.type = LLAMA_VOCAB_TYPE_NONE;
|
vocab.type = LLAMA_VOCAB_TYPE_NONE;
|
||||||
|
|
||||||
// default special tokens
|
// default special tokens
|
||||||
@ -4232,7 +4237,7 @@ static void llm_load_vocab(
|
|||||||
vocab.linefeed_id = -1;
|
vocab.linefeed_id = -1;
|
||||||
|
|
||||||
return;
|
return;
|
||||||
} else if (tokenizer_name == "llama") {
|
} else if (tokenizer_model == "llama") {
|
||||||
vocab.type = LLAMA_VOCAB_TYPE_SPM;
|
vocab.type = LLAMA_VOCAB_TYPE_SPM;
|
||||||
|
|
||||||
// default special tokens
|
// default special tokens
|
||||||
@ -4277,9 +4282,27 @@ static void llm_load_vocab(
|
|||||||
if (add_space_prefix_keyidx != -1) {
|
if (add_space_prefix_keyidx != -1) {
|
||||||
vocab.add_space_prefix = gguf_get_val_bool(ctx, add_space_prefix_keyidx);
|
vocab.add_space_prefix = gguf_get_val_bool(ctx, add_space_prefix_keyidx);
|
||||||
} // The default value of add_space_prefix is true.
|
} // The default value of add_space_prefix is true.
|
||||||
} else if (tokenizer_name == "gpt2") {
|
} else if (tokenizer_model == "bert") {
|
||||||
vocab.type = LLAMA_VOCAB_TYPE_BPE;
|
vocab.type = LLAMA_VOCAB_TYPE_WPM;
|
||||||
|
|
||||||
|
// default special tokens
|
||||||
|
vocab.special_bos_id = -1;
|
||||||
|
vocab.special_eos_id = -1;
|
||||||
|
vocab.special_unk_id = 100;
|
||||||
|
vocab.special_sep_id = 102;
|
||||||
|
vocab.special_pad_id = 0;
|
||||||
|
vocab.special_cls_id = 101;
|
||||||
|
vocab.special_mask_id = 103;
|
||||||
|
vocab.add_space_prefix = false;
|
||||||
|
} else {
|
||||||
|
if (tokenizer_model == "gpt2") {
|
||||||
|
vocab.type = LLAMA_VOCAB_TYPE_BPE;
|
||||||
|
} else {
|
||||||
|
LLAMA_LOG_WARN("%s: unknown tokenizer: '%s'", __func__, tokenizer_model.c_str());
|
||||||
|
LLAMA_LOG_WARN("%s: using default tokenizer: 'llama'", __func__);
|
||||||
|
vocab.type = LLAMA_VOCAB_TYPE_SPM;
|
||||||
|
return;
|
||||||
|
}
|
||||||
// read bpe merges and populate bpe ranks
|
// read bpe merges and populate bpe ranks
|
||||||
const int merges_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_MERGES).c_str());
|
const int merges_keyidx = gguf_find_key(ctx, kv(LLM_KV_TOKENIZER_MERGES).c_str());
|
||||||
if (merges_keyidx == -1) {
|
if (merges_keyidx == -1) {
|
||||||
@ -4313,23 +4336,50 @@ static void llm_load_vocab(
|
|||||||
vocab.special_pad_id = -1;
|
vocab.special_pad_id = -1;
|
||||||
vocab.special_cls_id = -1;
|
vocab.special_cls_id = -1;
|
||||||
vocab.special_mask_id = -1;
|
vocab.special_mask_id = -1;
|
||||||
} else if (tokenizer_name == "bert") {
|
}
|
||||||
vocab.type = LLAMA_VOCAB_TYPE_WPM;
|
|
||||||
|
|
||||||
// default special tokens
|
// for now, only BPE models have pre-tokenizers
|
||||||
vocab.special_bos_id = -1;
|
if (vocab.type == LLAMA_VOCAB_TYPE_BPE) {
|
||||||
vocab.special_eos_id = -1;
|
if (tokenizer_pre.empty()) {
|
||||||
vocab.special_unk_id = 100;
|
LLAMA_LOG_WARN("%s: missing pre-tokenizer type, using: 'default'\n", __func__);
|
||||||
vocab.special_sep_id = 102;
|
LLAMA_LOG_WARN("%s: \n", __func__);
|
||||||
vocab.special_pad_id = 0;
|
LLAMA_LOG_WARN("%s: ************************************ \n", __func__);
|
||||||
vocab.special_cls_id = 101;
|
LLAMA_LOG_WARN("%s: GENERATION QUALITY WILL BE DEGRADED! \n", __func__);
|
||||||
vocab.special_mask_id = 103;
|
LLAMA_LOG_WARN("%s: CONSIDER REGENERATING THE MODEL \n", __func__);
|
||||||
vocab.add_space_prefix = false;
|
LLAMA_LOG_WARN("%s: ************************************ \n", __func__);
|
||||||
|
LLAMA_LOG_WARN("%s: \n", __func__);
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "default") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "llama3" ||
|
||||||
|
tokenizer_pre == "llama-v3" ||
|
||||||
|
tokenizer_pre == "llama-bpe") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_LLAMA3;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "deepseek-llm") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "deepseek-coder") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "falcon") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_FALCON;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "mpt") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_MPT;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "starcoder") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_STARCODER;
|
||||||
|
} else if (
|
||||||
|
tokenizer_pre == "gpt-2") {
|
||||||
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_GPT2;
|
||||||
|
} else {
|
||||||
|
throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
|
||||||
|
}
|
||||||
} else {
|
} else {
|
||||||
LLAMA_LOG_WARN("%s: unknown tokenizer: '%s'", __func__, tokenizer_name.c_str());
|
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
|
||||||
LLAMA_LOG_WARN("%s: using default tokenizer: 'llama'", __func__);
|
|
||||||
|
|
||||||
vocab.type = LLAMA_VOCAB_TYPE_SPM;
|
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
@ -11834,7 +11884,7 @@ static uint8_t llama_token_to_byte(const llama_vocab& vocab, llama_token id) {
|
|||||||
}
|
}
|
||||||
case LLAMA_VOCAB_TYPE_BPE: {
|
case LLAMA_VOCAB_TYPE_BPE: {
|
||||||
GGML_ASSERT(false);
|
GGML_ASSERT(false);
|
||||||
return unicode_utf8_to_byte(token_data.text);
|
return unicode_utf8_to_byte(token_data.text); // TODO: why is this here after GGML_ASSERT?
|
||||||
}
|
}
|
||||||
case LLAMA_VOCAB_TYPE_WPM: {
|
case LLAMA_VOCAB_TYPE_WPM: {
|
||||||
GGML_ASSERT(false);
|
GGML_ASSERT(false);
|
||||||
@ -12056,7 +12106,79 @@ struct llm_tokenizer_bpe {
|
|||||||
|
|
||||||
void tokenize(const std::string & text, std::vector<llama_vocab::id> & output) {
|
void tokenize(const std::string & text, std::vector<llama_vocab::id> & output) {
|
||||||
int final_prev_index = -1;
|
int final_prev_index = -1;
|
||||||
auto word_collection = bpe_gpt2_preprocess(text);
|
|
||||||
|
std::vector<std::string> word_collection;
|
||||||
|
switch (vocab.type) {
|
||||||
|
case LLAMA_VOCAB_TYPE_BPE:
|
||||||
|
switch (vocab.type_pre) {
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_LLAMA3:
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
// original regex from tokenizer.json
|
||||||
|
//"(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
||||||
|
|
||||||
|
// adapted: https://github.com/ggerganov/llama.cpp/pull/6920#issuecomment-2080233989
|
||||||
|
"(?:'[sS]|'[tT]|'[rR][eE]|'[vV][eE]|'[mM]|'[lL][lL]|'[dD])|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM:
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"[\r\n]",
|
||||||
|
"\\s?[A-Za-zµÀ-ÖØ-öø-ƺƼ-ƿDŽ-ʓʕ-ʯͰ-ͳͶͷͻ-ͽͿΆΈ-ΊΌΎ-ΡΣ-ϵϷ-ҁҊ-ԯԱ-ՖႠ-ჅᎠ-Ᏽᏸ-ᏽᲐ-ᲺᲽ-Ჿᴀ-ᴫᵫ-ᵷᵹ-ᶚḀ-ἕἘ-Ἕἠ-ὅὈ-Ὅὐ-ὗὙὛὝὟ-ώᾀ-ᾴᾶ-ᾼιῂ-ῄῆ-ῌῐ-ΐῖ-Ίῠ-Ῥῲ-ῴῶ-ῼℂℇℊ-ℓℕℙ-ℝℤΩℨK-ℭℯ-ℴℹℼ-ℿⅅ-ⅉⅎↃↄⰀ-ⱻⱾ-ⳤⳫ-ⳮⳲⳳꙀ-ꙭꚀ-ꚛꜢ-ꝯꝱ-ꞇꞋ-ꞎꭰ-ꮿff-stﬓ-ﬗＡ-Ｚａ-ｚ𐐀-𐑏𐒰-𐓓𐓘-𐓻𐲀-𐲲𐳀-𐳲𑢠-𑣟𞤀-𞥃]+",
|
||||||
|
"\\s?[!-/:-~！-／：-～‘-‟ -。]+",
|
||||||
|
"\\s+$",
|
||||||
|
"[一-龥ࠀ-一가-]+",
|
||||||
|
"\\p{N}+",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER:
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"[\r\n]",
|
||||||
|
"\\s?\\p{L}+",
|
||||||
|
"\\s?\\p{P}+",
|
||||||
|
"[一-龥ࠀ-一가-]+",
|
||||||
|
"\\p{N}+",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_FALCON:
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"[\\p{P}\\$\\+<=>\\^~\\|]+",
|
||||||
|
"'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
|
||||||
|
"\\p{N}+",
|
||||||
|
"[0-9][0-9][0-9]",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_MPT:
|
||||||
|
// TODO: MPT pre-tokenization regexes are unknown
|
||||||
|
// the following are close, but not exact. run the following:
|
||||||
|
// ./bin/test-tokenizer-0 ../models/ggml-vocab-mpt.gguf
|
||||||
|
GGML_ASSERT("MPT pre-tokenization regexes are unknown - fixes needed");
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"\\s?\\p{L}+",
|
||||||
|
"\\s?\\p{P}+",
|
||||||
|
"'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_STARCODER:
|
||||||
|
case LLAMA_VOCAB_PRE_TYPE_GPT2:
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
default:
|
||||||
|
// default regex for BPE tokenization pre-processing
|
||||||
|
word_collection = unicode_regex_split(text, {
|
||||||
|
"[\\p{P}\\$\\+<=>\\^~\\|]+",
|
||||||
|
"'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
|
||||||
|
"\\p{N}+",
|
||||||
|
"[0-9][0-9][0-9]",
|
||||||
|
});
|
||||||
|
break;
|
||||||
|
}
|
||||||
|
break;
|
||||||
|
default:
|
||||||
|
GGML_ASSERT(false);
|
||||||
|
break;
|
||||||
|
}
|
||||||
|
|
||||||
symbols_final.clear();
|
symbols_final.clear();
|
||||||
|
|
||||||
@ -12183,145 +12305,6 @@ private:
|
|||||||
work_queue.push(bigram);
|
work_queue.push(bigram);
|
||||||
}
|
}
|
||||||
|
|
||||||
std::vector<std::string> bpe_gpt2_preprocess(const std::string & text) {
|
|
||||||
std::vector<std::string> bpe_words;
|
|
||||||
std::vector<std::string> bpe_encoded_words;
|
|
||||||
|
|
||||||
std::string token = "";
|
|
||||||
// GPT2 system regex: 's|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+
|
|
||||||
bool collecting_numeric = false;
|
|
||||||
bool collecting_letter = false;
|
|
||||||
bool collecting_special = false;
|
|
||||||
bool collecting_whitespace_lookahead = false;
|
|
||||||
bool collecting = false;
|
|
||||||
|
|
||||||
std::vector<std::string> text_utf;
|
|
||||||
text_utf.reserve(text.size());
|
|
||||||
bpe_words.reserve(text.size());
|
|
||||||
bpe_encoded_words.reserve(text.size());
|
|
||||||
|
|
||||||
const auto cpts = unicode_cpts_from_utf8(text);
|
|
||||||
for (size_t i = 0; i < cpts.size(); ++i)
|
|
||||||
text_utf.emplace_back(unicode_cpt_to_utf8(cpts[i]));
|
|
||||||
|
|
||||||
for (int i = 0; i < (int)text_utf.size(); i++) {
|
|
||||||
const std::string & utf_char = text_utf[i];
|
|
||||||
bool split_condition = false;
|
|
||||||
int bytes_remain = text_utf.size() - i;
|
|
||||||
// forward backward lookups
|
|
||||||
const std::string & utf_char_next = (i + 1 < (int)text_utf.size()) ? text_utf[i + 1] : "";
|
|
||||||
const std::string & utf_char_next_next = (i + 2 < (int)text_utf.size()) ? text_utf[i + 2] : "";
|
|
||||||
|
|
||||||
// handling contractions
|
|
||||||
if (!split_condition && bytes_remain >= 2) {
|
|
||||||
// 's|'t|'m|'d
|
|
||||||
if (utf_char == "\'" && (utf_char_next == "s" || utf_char_next == "t" || utf_char_next == "m" || utf_char_next == "d")) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
if (split_condition) {
|
|
||||||
if (token.size()) {
|
|
||||||
bpe_words.emplace_back(token); // push previous content as token
|
|
||||||
}
|
|
||||||
token = utf_char + utf_char_next;
|
|
||||||
bpe_words.emplace_back(token);
|
|
||||||
token = "";
|
|
||||||
i++;
|
|
||||||
continue;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
if (!split_condition && bytes_remain >= 3) {
|
|
||||||
// 're|'ve|'ll
|
|
||||||
if (utf_char == "\'" && (
|
|
||||||
(utf_char_next == "r" && utf_char_next_next == "e") ||
|
|
||||||
(utf_char_next == "v" && utf_char_next_next == "e") ||
|
|
||||||
(utf_char_next == "l" && utf_char_next_next == "l"))
|
|
||||||
) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
if (split_condition) {
|
|
||||||
// current token + next token can be defined
|
|
||||||
if (token.size()) {
|
|
||||||
bpe_words.emplace_back(token); // push previous content as token
|
|
||||||
}
|
|
||||||
token = utf_char + utf_char_next + utf_char_next_next;
|
|
||||||
bpe_words.emplace_back(token); // the contraction
|
|
||||||
token = "";
|
|
||||||
i += 2;
|
|
||||||
continue;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!split_condition && !collecting) {
|
|
||||||
if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_LETTER || (!token.size() && utf_char == " " && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_LETTER)) {
|
|
||||||
collecting_letter = true;
|
|
||||||
collecting = true;
|
|
||||||
}
|
|
||||||
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_DIGIT || (!token.size() && utf_char == " " && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_DIGIT)) {
|
|
||||||
collecting_numeric = true;
|
|
||||||
collecting = true;
|
|
||||||
}
|
|
||||||
else if (
|
|
||||||
((unicode_cpt_type(utf_char) != CODEPOINT_TYPE_LETTER && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_DIGIT) && (unicode_cpt_type(utf_char) != CODEPOINT_TYPE_WHITESPACE)) ||
|
|
||||||
(!token.size() && utf_char == " " && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_LETTER && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_DIGIT && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_WHITESPACE)
|
|
||||||
) {
|
|
||||||
collecting_special = true;
|
|
||||||
collecting = true;
|
|
||||||
}
|
|
||||||
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_WHITESPACE) {
|
|
||||||
collecting_whitespace_lookahead = true;
|
|
||||||
collecting = true;
|
|
||||||
}
|
|
||||||
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
else if (!split_condition && collecting) {
|
|
||||||
if (collecting_letter && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_LETTER) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
else if (collecting_numeric && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_DIGIT) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
else if (collecting_special && (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_LETTER || unicode_cpt_type(utf_char) == CODEPOINT_TYPE_DIGIT || unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE)) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
else if (collecting_whitespace_lookahead && (unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_LETTER || unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_DIGIT)) {
|
|
||||||
split_condition = true;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (utf_char_next == "") {
|
|
||||||
split_condition = true; // final
|
|
||||||
token += utf_char;
|
|
||||||
}
|
|
||||||
|
|
||||||
if (split_condition) {
|
|
||||||
if (token.size()) {
|
|
||||||
bpe_words.emplace_back(token);
|
|
||||||
}
|
|
||||||
token = utf_char;
|
|
||||||
collecting = false;
|
|
||||||
collecting_letter = false;
|
|
||||||
collecting_numeric = false;
|
|
||||||
collecting_special = false;
|
|
||||||
collecting_whitespace_lookahead = false;
|
|
||||||
}
|
|
||||||
else {
|
|
||||||
token += utf_char;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
for (std::string & word : bpe_words) {
|
|
||||||
std::string encoded_token = "";
|
|
||||||
for (char & c : word) {
|
|
||||||
encoded_token += unicode_byte_to_utf8(c);
|
|
||||||
}
|
|
||||||
bpe_encoded_words.emplace_back(encoded_token);
|
|
||||||
}
|
|
||||||
|
|
||||||
return bpe_encoded_words;
|
|
||||||
}
|
|
||||||
|
|
||||||
const llama_vocab & vocab;
|
const llama_vocab & vocab;
|
||||||
|
|
||||||
std::vector<llm_symbol> symbols;
|
std::vector<llm_symbol> symbols;
|
||||||
@ -12641,7 +12624,7 @@ static std::vector<llama_vocab::id> llama_tokenize_internal(const llama_vocab &
|
|||||||
} break;
|
} break;
|
||||||
case LLAMA_VOCAB_TYPE_BPE:
|
case LLAMA_VOCAB_TYPE_BPE:
|
||||||
{
|
{
|
||||||
if (add_special && vocab.special_add_bos == 1) {
|
if (add_special && vocab.special_add_bos != 0) {
|
||||||
GGML_ASSERT(vocab.special_bos_id != -1);
|
GGML_ASSERT(vocab.special_bos_id != -1);
|
||||||
output.push_back(vocab.special_bos_id);
|
output.push_back(vocab.special_bos_id);
|
||||||
}
|
}
|
||||||
|
|||||||
12
llama.h
12
llama.h
@ -69,6 +69,18 @@ extern "C" {
|
|||||||
LLAMA_VOCAB_TYPE_WPM = 3, // BERT tokenizer based on WordPiece
|
LLAMA_VOCAB_TYPE_WPM = 3, // BERT tokenizer based on WordPiece
|
||||||
};
|
};
|
||||||
|
|
||||||
|
// pre-tokenization types
|
||||||
|
enum llama_vocab_pre_type {
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_DEFAULT = 0,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_LLAMA3 = 1,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM = 2,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER = 3,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_FALCON = 4,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_MPT = 5,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_STARCODER = 6,
|
||||||
|
LLAMA_VOCAB_PRE_TYPE_GPT2 = 7,
|
||||||
|
};
|
||||||
|
|
||||||
// note: these values should be synchronized with ggml_rope
|
// note: these values should be synchronized with ggml_rope
|
||||||
// TODO: maybe move this enum to ggml.h (ggml_rope_type)
|
// TODO: maybe move this enum to ggml.h (ggml_rope_type)
|
||||||
enum llama_rope_type {
|
enum llama_rope_type {
|
||||||
|
|||||||
BIN
models/ggml-vocab-bert-bge.gguf
Normal file
BIN
models/ggml-vocab-bert-bge.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-bert-bge.gguf.inp
Normal file
102
models/ggml-vocab-bert-bge.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-bert-bge.gguf.out
Normal file
41
models/ggml-vocab-bert-bge.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
7592 2088
|
||||||
|
7592 2088
|
||||||
|
7592 2088
|
||||||
|
7592 2088
|
||||||
|
7592 2088 999
|
||||||
|
7592 1010 2088 999
|
||||||
|
7592 1010 2088 999
|
||||||
|
2023 2003 100 1012 18133 2361
|
||||||
|
1059 2692 18139 1021 8525 28418 2243 16233 20952 6979
|
||||||
|
1192 15290 29754 14150 1192 10260 1181 29755 29436 29741 10260 16856 29747 23925 10325
|
||||||
|
100
|
||||||
|
100 1006 3671 1007 100 1006 3674 7861 29147 2483 9530 16280 23854 1007 100 1006 2069 7861 29147 2072 2008 2038 2049 2219 19204 1007
|
||||||
|
7592
|
||||||
|
7592
|
||||||
|
7592
|
||||||
|
7592
|
||||||
|
7592
|
||||||
|
7592 7592
|
||||||
|
1006
|
||||||
|
1027
|
||||||
|
1005 3690
|
||||||
|
7592 1010 1061 1005 2035 999 2129 2024 2017 100 1029 1855 100 100 6207 100 100 14677 23632 22203 1811 1995
|
||||||
|
1017
|
||||||
|
3943
|
||||||
|
21211
|
||||||
|
21211 2509
|
||||||
|
21211 22394
|
||||||
|
21211 22394 2509
|
||||||
|
21211 22394 22394
|
||||||
|
21211 22394 22394 2509
|
||||||
|
21211 22394 22394 22394
|
||||||
|
100 1006 3671 1007 100 1006 3674 7861 29147 2483 9530 16280 23854 1007 100 100 1017 3943 21211 21211 2509 21211 22394 21211 22394 2509 21211 22394 22394 21211 22394 22394 2509 1017 1012 1017 1017 1012 1012 1017 1017 1012 1012 1012 1017 100 1029 1855 100 100 6207 100 100 14677 23632 22203 1811 1995 1011 1011 1011 1011 1011 1011 1027 1027 1027 1027 1027 1027 1027 1192 15290 29754 14150 1192 10260 1181 29755 29436 29741 10260 16856 29747 23925 10325 1005 1005 1005 1005 1005 1005 1036 1036 1036 1036 1036 1036 1036 1000 1000 1000 1000 1012 1012 1012 1012 1012 1012 999 999 999 999 999 999 1029 1029 1029 1029 1029 1029 1045 1005 2310 2042 1005 2409 2002 1005 1055 2045 1010 1005 2128 2017 2469 1029 1005 1049 2025 2469 1045 1005 2222 2191 2009 1010 1005 1040 2017 2066 2070 5572 1029 2057 1005 2310 1037 1005 2222
|
||||||
BIN
models/ggml-vocab-deepseek-coder.gguf
Normal file
BIN
models/ggml-vocab-deepseek-coder.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-deepseek-coder.gguf.inp
Normal file
102
models/ggml-vocab-deepseek-coder.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-deepseek-coder.gguf.out
Normal file
41
models/ggml-vocab-deepseek-coder.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
207
|
||||||
|
243
|
||||||
|
315
|
||||||
|
184
|
||||||
|
185
|
||||||
|
185 185
|
||||||
|
185 185 185
|
||||||
|
184 185
|
||||||
|
17535 1835
|
||||||
|
414 9489 1835
|
||||||
|
17535 5414
|
||||||
|
414 9489 5414
|
||||||
|
414 9489 5414 0
|
||||||
|
17535 11 1835 0
|
||||||
|
414 9489 11 1835 0
|
||||||
|
437 317 12394 99 234 13 14789
|
||||||
|
86 15 19 23 207 22 83 3963 27659 26078 3934 14072
|
||||||
|
1593 6478 616 2251 14994
|
||||||
|
155 239 209 155 239 114 155 239 228 155 240 220 155 239 224 155 240 211 155 239 231 155 239 115 155 239 240 155 240 210 155 239 240 155 239 95 155 239 114 155 239 214 155 239 210 155 239 236 155 239 214 155 240 210 155 239 218
|
||||||
|
10047 235 209 334 8760 8 12394 233 114 350 222 10047 221 104 169 116 224 334 4684 3909 992 24330 262 29651 612 8 207 156 237 214 334 5950 992 78 12896 344 638 891 1372 10736 8
|
||||||
|
17535
|
||||||
|
414 9489
|
||||||
|
207 414 9489
|
||||||
|
243 414 9489
|
||||||
|
315 414 9489
|
||||||
|
315 414 9489 185 315 414 9489
|
||||||
|
334
|
||||||
|
185 405
|
||||||
|
6 2895
|
||||||
|
17535 11 320 6 435 0 1717 417 340 12394 233 210 3015 19100 608 9413 2668 16 18 16 19 16 20 16 1393 169 121 239
|
||||||
|
18
|
||||||
|
18 18
|
||||||
|
18 18 18
|
||||||
|
18 18 18 18
|
||||||
|
18 18 18 18 18
|
||||||
|
18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18 18 18
|
||||||
|
185 207 185 185 207 185 185 185 207 12405 459 22758 185 243 185 315 185 251 185 730 185 10047 235 209 334 8760 8 12394 233 114 350 222 10047 221 104 169 116 224 334 4684 3909 992 24330 262 29651 612 8 207 156 237 214 12394 99 234 10047 99 234 207 18 207 18 18 207 18 18 18 207 18 18 18 18 207 18 18 18 18 18 207 18 18 18 18 18 18 207 18 18 18 18 18 18 18 207 18 18 18 18 18 18 18 18 207 18 13 18 207 18 524 18 207 18 1202 18 207 155 239 209 155 239 114 155 239 228 155 240 220 155 239 224 155 240 211 155 239 231 155 239 115 155 239 240 155 240 210 155 239 240 155 239 95 155 239 114 155 239 214 10047 233 210 3015 19100 608 9413 2668 16 18 16 19 16 20 16 1393 169 121 239 18155 374 17194 28 2861 6478 616 2251 14994 31269 4191 6 4686 4686 10252 3358 3358 3409 524 15330 3023 15031 5668 303 6 312 798 651 83 839 362 6 82 741 11 651 1369 340 2037 30 651 44 441 2037 303 6 642 1098 359 11 651 35 340 833 738 10860 30 998 6 10709 245 6 75 43
|
||||||
BIN
models/ggml-vocab-deepseek-llm.gguf
Normal file
BIN
models/ggml-vocab-deepseek-llm.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-deepseek-llm.gguf.inp
Normal file
102
models/ggml-vocab-deepseek-llm.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-deepseek-llm.gguf.out
Normal file
41
models/ggml-vocab-deepseek-llm.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
207
|
||||||
|
243
|
||||||
|
300
|
||||||
|
184
|
||||||
|
185
|
||||||
|
185 185
|
||||||
|
185 185 185
|
||||||
|
184 185
|
||||||
|
17464 1843
|
||||||
|
37727 1843
|
||||||
|
17464 5427
|
||||||
|
37727 5427
|
||||||
|
37727 5427 0
|
||||||
|
17464 11 1843 0
|
||||||
|
37727 11 1843 0
|
||||||
|
437 317 12356 99 234 13 14743
|
||||||
|
86 15 19 23 207 22 83 3970 27519 26016 3944 14025
|
||||||
|
1603 6476 620 91754
|
||||||
|
71374 209 71374 114 71374 228 155 240 220 71374 224 155 240 211 71374 231 71374 115 71374 240 155 240 210 71374 240 71374 95 71374 114 71374 214 71374 210 71374 236 71374 214 155 240 210 71374 218
|
||||||
|
10044 95300 334 8754 8 33701 114 350 222 10044 221 104 46713 334 34732 996 24250 262 80923 8 207 37103 214 334 5956 89213 344 643 895 1377 10728 8
|
||||||
|
17464
|
||||||
|
37727
|
||||||
|
207 37727
|
||||||
|
243 37727
|
||||||
|
300 37727
|
||||||
|
300 37727 185 300 37727
|
||||||
|
334
|
||||||
|
185 403
|
||||||
|
6 2906
|
||||||
|
17464 11 320 6 436 0 1724 418 340 33701 210 3025 19017 612 9407 2681 16 18 16 19 16 20 16 1398 68940 239
|
||||||
|
18
|
||||||
|
18 18
|
||||||
|
18 18 18
|
||||||
|
18 18 18 18
|
||||||
|
18 18 18 18 18
|
||||||
|
18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18 18
|
||||||
|
18 18 18 18 18 18 18 18 18
|
||||||
|
185 207 185 185 207 185 185 185 207 11969 486 22504 185 243 185 300 185 251 185 663 185 10044 95300 334 8754 8 33701 114 350 222 10044 221 104 46713 334 34732 996 24250 262 80923 8 207 37103 214 12356 99 234 10044 99 234 207 18 207 18 18 207 18 18 18 207 18 18 18 18 207 18 18 18 18 18 207 18 18 18 18 18 18 207 18 18 18 18 18 18 18 207 18 18 18 18 18 18 18 18 207 18 13 18 207 18 526 18 207 18 1204 18 207 71374 209 71374 114 71374 228 155 240 220 71374 224 155 240 211 71374 231 71374 115 71374 240 155 240 210 71374 240 71374 95 71374 114 71374 214 71899 210 3025 19017 612 9407 2681 16 18 16 19 16 20 16 1398 68940 239 78827 55170 76659 620 91754 31116 36804 4885 4885 10897 4390 4390 41047 15278 3033 14986 5675 304 6 313 803 655 33326 362 6 82 745 11 655 1374 340 2049 30 655 44 441 2049 304 6 647 1099 359 11 655 35 340 837 742 10842 30 1003 6 10699 245 6 75 43
|
||||||
Binary file not shown.
102
models/ggml-vocab-falcon.gguf.inp
Normal file
102
models/ggml-vocab-falcon.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-falcon.gguf.out
Normal file
41
models/ggml-vocab-falcon.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
204
|
||||||
|
258
|
||||||
|
466
|
||||||
|
192
|
||||||
|
193
|
||||||
|
1001
|
||||||
|
11331
|
||||||
|
19125
|
||||||
|
9856 1079
|
||||||
|
23090 1079
|
||||||
|
9856 2889
|
||||||
|
23090 2889
|
||||||
|
23090 2889 12
|
||||||
|
9856 23 1079 12
|
||||||
|
23090 23 1079 12
|
||||||
|
414 304 3346 111 231 25 29247
|
||||||
|
98 55866 204 34 16682 7149 36190 6869 11481
|
||||||
|
150 133 6207 151 215 150 134 5052 133 6279 5052 223 151 216 49679 123 53110 47043 7795
|
||||||
|
38154 206 38154 126 38154 225 167 237 217 38154 221 167 237 208 38154 228 38154 127 38154 237 167 237 207 38154 237 38154 107 38154 126 38154 211 38154 207 38154 233 38154 211 167 237 207 38154 215
|
||||||
|
2571 232 206 204 19 11003 20 8196 126 283 219 48778 116 13392 204 19 51831 732 63209 1741 7955 522 20 22438 211 204 19 7927 53360 325 504 701 946 10930 20
|
||||||
|
9856
|
||||||
|
23090
|
||||||
|
204 23090
|
||||||
|
258 23090
|
||||||
|
466 23090
|
||||||
|
466 23090 742 23090
|
||||||
|
204 19
|
||||||
|
1212 40
|
||||||
|
18 4932
|
||||||
|
9856 23 291 18 436 12 1265 362 299 8196 207 204 42 50087 123 2727 20300 32022 133 234 17419 30137 28 7858 181 133 236
|
||||||
|
30
|
||||||
|
3138
|
||||||
|
22287
|
||||||
|
22287 30
|
||||||
|
22287 3138
|
||||||
|
22287 22287
|
||||||
|
22287 22287 30
|
||||||
|
22287 22287 3138
|
||||||
|
22287 22287 22287
|
||||||
|
1212 4824 1001 1212 192 204 663 49453 2069 742 561 1501 193 2571 232 206 204 19 11003 20 8196 126 283 219 48778 116 13392 204 19 51831 732 63209 1741 7955 522 20 22438 211 3346 111 231 2571 111 231 204 30 204 3138 204 22287 204 22287 30 204 22287 3138 204 22287 22287 204 22287 22287 30 204 22287 22287 3138 204 30 25 30 204 30 513 30 204 30 951 30 27171 236 206 38154 126 38154 225 167 237 217 38154 221 167 237 208 38154 228 38154 127 38154 237 167 237 207 38154 237 38154 107 38154 126 38154 211 20589 207 204 42 50087 123 2727 20300 32022 133 234 17419 30137 28 7858 181 133 236 204 37057 2228 10666 5052 133 6207 151 215 150 134 5052 133 6279 5052 223 151 216 49679 123 53110 47043 7795 204 7544 7544 7544 8543 8543 17593 3513 3513 12844 51520 17664 4247 295 18 298 650 204 18 95 693 332 18 94 629 23 204 18 1553 299 1310 42 204 18 56 416 1310 295 18 567 717 334 23 204 18 47 299 606 596 6696 42 703 18 16139 241 18 87 55
|
||||||
BIN
models/ggml-vocab-gpt-2.gguf
Normal file
BIN
models/ggml-vocab-gpt-2.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-gpt-2.gguf.inp
Normal file
102
models/ggml-vocab-gpt-2.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-gpt-2.gguf.out
Normal file
41
models/ggml-vocab-gpt-2.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
220
|
||||||
|
220 220
|
||||||
|
220 220 220
|
||||||
|
197
|
||||||
|
198
|
||||||
|
628
|
||||||
|
628 198
|
||||||
|
197 198
|
||||||
|
15496 995
|
||||||
|
18435 995
|
||||||
|
15496 2159
|
||||||
|
18435 2159
|
||||||
|
18435 2159 0
|
||||||
|
15496 11 995 0
|
||||||
|
18435 11 995 0
|
||||||
|
428 318 12520 99 247 13 20322
|
||||||
|
86 47202 767 28047 45961 288 82 7568 13415
|
||||||
|
22177 16843 141 231 15166 12466 121 16142 12466 239 141 232 30143 140 111 16142 21169 21727 31583 18849
|
||||||
|
157 252 222 157 252 114 157 252 241 157 253 233 157 252 237 157 253 224 157 252 244 157 252 115 157 252 253 157 253 223 157 252 253 157 252 95 157 252 114 157 252 227 157 252 223 157 252 249 157 252 227 157 253 223 157 252 231
|
||||||
|
8582 248 222 357 11265 8 30325 114 447 235 8582 234 104 37929 357 48101 795 13210 271 1673 36686 515 8 14519 227 357 8807 44805 326 468 663 898 11241 8
|
||||||
|
15496
|
||||||
|
18435
|
||||||
|
220 18435
|
||||||
|
220 220 18435
|
||||||
|
220 220 220 18435
|
||||||
|
220 220 220 18435 198 220 220 220 18435
|
||||||
|
357
|
||||||
|
198 796
|
||||||
|
6 6980
|
||||||
|
15496 11 331 6 439 0 1374 389 345 30325 223 5633 22755 239 46349 111 28839 101 18040 32432 98 43291 1485 1415 24309 25465 171 121 252
|
||||||
|
18
|
||||||
|
2091
|
||||||
|
20370
|
||||||
|
24840
|
||||||
|
2091 20370
|
||||||
|
24840 2091
|
||||||
|
24840 20370
|
||||||
|
24840 24840
|
||||||
|
24840 2091 20370
|
||||||
|
198 220 628 220 628 198 220 197 220 197 197 220 197 198 220 220 198 220 220 220 198 220 220 220 220 198 220 220 220 220 220 198 8582 248 222 357 11265 8 30325 114 447 235 8582 234 104 37929 357 48101 795 13210 271 1673 36686 515 8 14519 227 12520 99 247 8582 99 247 513 4747 23460 513 20370 23460 2091 23460 20370 23460 24840 23460 2091 20370 513 13 18 513 492 18 513 986 18 28053 252 222 157 252 114 157 252 241 157 253 233 157 252 237 157 253 224 157 252 244 157 252 115 157 252 253 157 253 223 157 252 253 157 252 95 157 252 114 157 252 227 47249 223 5633 22755 239 46349 111 28839 101 18040 32432 98 43291 1485 1415 24309 25465 171 121 252 40103 1421 18604 12466 121 16843 141 231 15166 12466 121 16142 12466 239 141 232 30143 140 111 16142 21169 21727 31583 18849 705 39115 6 33153 15506 63 15931 15931 16317 13896 3228 9805 3548 314 1053 587 705 44040 339 338 612 11 705 2200 345 1654 30 705 44 407 1654 314 1183 787 340 11 705 35 345 588 617 8887 30 775 6 26979 257 6 75 43
|
||||||
BIN
models/ggml-vocab-llama-bpe.gguf
Normal file
BIN
models/ggml-vocab-llama-bpe.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-llama-bpe.gguf.inp
Normal file
102
models/ggml-vocab-llama-bpe.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-llama-bpe.gguf.out
Normal file
41
models/ggml-vocab-llama-bpe.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
220
|
||||||
|
256
|
||||||
|
262
|
||||||
|
197
|
||||||
|
198
|
||||||
|
271
|
||||||
|
1432
|
||||||
|
1602
|
||||||
|
9906 1917
|
||||||
|
22691 1917
|
||||||
|
9906 4435
|
||||||
|
22691 4435
|
||||||
|
22691 4435 0
|
||||||
|
9906 11 1917 0
|
||||||
|
22691 11 1917 0
|
||||||
|
420 374 11410 99 247 13 11055
|
||||||
|
86 23904 220 22 83 2005 42908 11729 3013 17156
|
||||||
|
79862 102118 13373 64571 34694 3114 112203 80112
|
||||||
|
21549 222 98629 241 45358 233 21549 237 45358 224 21549 244 21549 115 21549 253 45358 223 21549 253 21549 95 98629 227 21549 223 21549 249 21549 227 45358 223 21549 231
|
||||||
|
9468 248 222 320 8416 8 27623 114 102470 9468 234 104 31643 320 36773 100166 98634 8 26602 227 320 3323 43465 430 706 1202 1866 4037 8
|
||||||
|
9906
|
||||||
|
22691
|
||||||
|
220 22691
|
||||||
|
256 22691
|
||||||
|
262 22691
|
||||||
|
262 22691 198 262 22691
|
||||||
|
320
|
||||||
|
198 284
|
||||||
|
6 11639
|
||||||
|
9906 11 379 65948 0 2650 527 499 27623 223 949 37046 101067 19000 23182 102301 9263 18136 16 36827 21909
|
||||||
|
18
|
||||||
|
1644
|
||||||
|
8765
|
||||||
|
8765 18
|
||||||
|
8765 1644
|
||||||
|
8765 8765
|
||||||
|
8765 8765 18
|
||||||
|
8765 8765 1644
|
||||||
|
8765 8765 8765
|
||||||
|
198 4815 15073 66597 8004 1602 2355 79772 11187 9468 248 222 320 8416 8 27623 114 102470 9468 234 104 31643 320 36773 100166 98634 8 26602 227 11410 99 247 9468 99 247 220 18 220 1644 220 8765 220 8765 18 220 8765 1644 220 8765 8765 220 8765 8765 18 220 8765 8765 1644 220 18 13 18 220 18 497 18 220 18 1131 18 220 21549 222 98629 241 45358 233 21549 237 45358 224 21549 244 21549 115 21549 253 45358 223 21549 253 21549 95 98629 227 76460 223 949 37046 101067 19000 23182 102301 9263 18136 16 36827 21909 56560 54337 19175 102118 13373 64571 34694 3114 112203 80112 3436 106451 14196 14196 74694 3089 3089 29249 17523 3001 27708 7801 358 3077 1027 364 83 820 568 596 1070 11 364 793 499 2771 30 364 44 539 2771 358 3358 1304 433 11 364 35 499 1093 1063 15600 30 1226 6 43712 264 64966 43
|
||||||
Binary file not shown.
102
models/ggml-vocab-llama-spm.gguf.inp
Normal file
102
models/ggml-vocab-llama-spm.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-llama-spm.gguf.out
Normal file
41
models/ggml-vocab-llama-spm.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
259
|
||||||
|
1678
|
||||||
|
268
|
||||||
|
29871 12
|
||||||
|
29871 13
|
||||||
|
29871 13 13
|
||||||
|
29871 13 13 13
|
||||||
|
29871 12 13
|
||||||
|
15043 3186
|
||||||
|
29871 15043 3186
|
||||||
|
15043 2787
|
||||||
|
29871 15043 2787
|
||||||
|
29871 15043 2787 29991
|
||||||
|
15043 29892 3186 29991
|
||||||
|
29871 15043 29892 3186 29991
|
||||||
|
29871 445 338 29871 243 162 169 156 29889 8223
|
||||||
|
281 29900 29946 29947 29871 29955 9161 13535 18031 2176 6905
|
||||||
|
1538 4851 665 1386 29713 1305
|
||||||
|
29871 31849 31324 31934 228 162 142 228 161 146 228 162 133 228 161 153 228 161 186 31708 228 162 132 31708 228 161 165 31324 228 161 136 228 161 132 228 161 158 228 161 136 228 162 132 228 161 140
|
||||||
|
29871 243 162 157 131 313 8945 29897 29871 243 162 155 185 30722 243 162 143 174 30598 313 20787 953 3848 275 16125 630 29897 29871 31681 313 6194 953 29877 2397 393 756 967 1914 5993 29897
|
||||||
|
15043
|
||||||
|
29871 15043
|
||||||
|
259 15043
|
||||||
|
1678 15043
|
||||||
|
268 15043
|
||||||
|
268 15043 13 1678 15043
|
||||||
|
29871 313
|
||||||
|
29871 13 353
|
||||||
|
525 3152
|
||||||
|
15043 29892 343 29915 497 29991 1128 526 366 29871 243 162 155 132 1577 30672 31522 30505 11548 31041 30732 29896 29941 29896 29946 29896 29945 29896 30408 30739
|
||||||
|
29871 29941
|
||||||
|
29871 29941 29941
|
||||||
|
29871 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 13 29871 13 13 29871 13 13 13 29871 12 29871 12 12 29871 12 13 259 13 1678 13 268 13 418 13 243 162 157 131 313 8945 29897 29871 243 162 155 185 30722 243 162 143 174 30598 313 20787 953 3848 275 16125 630 29897 29871 31681 29871 243 162 169 156 243 162 169 156 29871 29941 29871 29941 29941 29871 29941 29941 29941 29871 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29941 29941 29871 29941 29889 29941 29871 29941 636 29941 29871 29941 856 29941 29871 31849 31324 31934 228 162 142 228 161 146 228 162 133 228 161 153 228 161 186 31708 228 162 132 31708 228 161 165 31324 228 161 136 243 162 155 132 1577 30672 31522 30505 11548 31041 30732 29896 29941 29896 29946 29896 29945 29896 30408 30739 448 23648 2751 25512 1538 4851 665 1386 29713 1305 14550 4907 11120 16159 16159 16159 15945 15945 3045 636 6824 6824 6824 8773 8773 8773 306 29915 345 1063 525 29873 1025 540 29915 29879 727 29892 525 1525 366 1854 29973 525 29924 451 1854 306 29915 645 1207 372 29892 525 29928 366 763 777 23429 29973 1334 29915 29963 29872 263 29915 29880 29931
|
||||||
Binary file not shown.
102
models/ggml-vocab-mpt.gguf.inp
Normal file
102
models/ggml-vocab-mpt.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-mpt.gguf.out
Normal file
41
models/ggml-vocab-mpt.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
209
|
||||||
|
50276
|
||||||
|
50275
|
||||||
|
186
|
||||||
|
187
|
||||||
|
535
|
||||||
|
2756
|
||||||
|
186 187
|
||||||
|
12092 1533
|
||||||
|
24387 1533
|
||||||
|
12092 3645
|
||||||
|
24387 3645
|
||||||
|
24387 3645 2
|
||||||
|
12092 13 1533 2
|
||||||
|
24387 13 1533 2
|
||||||
|
436 310 22692 101 236 15 14161
|
||||||
|
88 27244 818 16853 16392 20505 4989 11917
|
||||||
|
32520 11514 1068 8713 38177 13396 3415 9925 12559 10453 1389
|
||||||
|
18081 211 18081 116 18081 230 39936 222 18081 226 39936 213 18081 233 18081 117 18081 242 39936 212 18081 242 18081 97 18081 116 18081 216 18081 212 18081 238 18081 216 39936 212 18081 220
|
||||||
|
14931 237 211 313 6320 10 49042 116 325 224 14931 223 106 171 118 226 313 34263 802 13511 261 32147 456 10 3384 239 216 313 7483 802 80 8020 326 556 697 1211 10669 10
|
||||||
|
12092
|
||||||
|
24387
|
||||||
|
50276 12092
|
||||||
|
50275 12092
|
||||||
|
50274 12092
|
||||||
|
50274 12092 187 50274 12092
|
||||||
|
313
|
||||||
|
187 426
|
||||||
|
8 8685
|
||||||
|
12092 13 340 8 455 2 1359 403 368 49042 212 3736 15367 41197 13610 19934 41869 21275 1012 1047 18795 40120 20422 241
|
||||||
|
20
|
||||||
|
1610
|
||||||
|
20084
|
||||||
|
26409
|
||||||
|
1610 20084
|
||||||
|
26409 1610
|
||||||
|
26409 20084
|
||||||
|
26409 26409
|
||||||
|
26409 1610 20084
|
||||||
|
586 1744 33525 186 209 623 28910 187 50276 187 50275 187 50274 187 50273 187 14931 237 211 313 6320 10 49042 116 325 224 14931 223 106 171 118 226 313 34263 802 13511 261 32147 456 10 3384 239 216 22692 101 236 14931 101 236 495 5922 30057 495 20084 495 26409 30057 20084 495 26409 1610 495 26409 20084 495 15 20 495 537 20 495 1051 20 209 18081 211 18081 116 18081 230 39936 222 18081 226 39936 213 18081 233 18081 117 18081 242 39936 212 18081 242 18081 97 18081 116 18081 216 14931 235 212 3736 15367 41197 13610 19934 41869 21275 1012 1047 18795 40120 20422 241 16081 6877 12880 11514 1068 8713 38177 13396 3415 9925 12559 10453 1389 42011 35033 34842 11202 9739 9739 33021 18963 4672 25561 8220 309 1849 644 686 42618 344 434 627 13 686 1848 368 2119 32 686 46 417 2119 309 1833 1056 352 13 686 37 368 751 690 10331 32 844 8 31516 247 8 77 45
|
||||||
BIN
models/ggml-vocab-phi-3.gguf
Normal file
BIN
models/ggml-vocab-phi-3.gguf
Normal file
Binary file not shown.
102
models/ggml-vocab-phi-3.gguf.inp
Normal file
102
models/ggml-vocab-phi-3.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-phi-3.gguf.out
Normal file
41
models/ggml-vocab-phi-3.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
259
|
||||||
|
1678
|
||||||
|
268
|
||||||
|
29871 12
|
||||||
|
29871 13
|
||||||
|
29871 13 13
|
||||||
|
29871 13 13 13
|
||||||
|
29871 12 13
|
||||||
|
15043 3186
|
||||||
|
29871 15043 3186
|
||||||
|
15043 2787
|
||||||
|
29871 15043 2787
|
||||||
|
29871 15043 2787 29991
|
||||||
|
15043 29892 3186 29991
|
||||||
|
29871 15043 29892 3186 29991
|
||||||
|
29871 445 338 29871 243 162 169 156 29889 8223
|
||||||
|
281 29900 29946 29947 29871 29955 9161 13535 18031 2176 6905
|
||||||
|
1538 4851 665 1386 29713 1305
|
||||||
|
29871 31849 31324 31934 228 162 142 228 161 146 228 162 133 228 161 153 228 161 186 31708 228 162 132 31708 228 161 165 31324 228 161 136 228 161 132 228 161 158 228 161 136 228 162 132 228 161 140
|
||||||
|
29871 243 162 157 131 313 8945 29897 29871 243 162 155 185 30722 243 162 143 174 30598 313 20787 953 3848 275 16125 630 29897 29871 31681 313 6194 953 29877 2397 393 756 967 1914 5993 29897
|
||||||
|
15043
|
||||||
|
29871 15043
|
||||||
|
259 15043
|
||||||
|
1678 15043
|
||||||
|
268 15043
|
||||||
|
268 15043 13 1678 15043
|
||||||
|
29871 313
|
||||||
|
29871 13 353
|
||||||
|
525 3152
|
||||||
|
15043 29892 343 29915 497 29991 1128 526 366 29871 243 162 155 132 1577 30672 31522 30505 11548 31041 30732 29896 29941 29896 29946 29896 29945 29896 30408 30739
|
||||||
|
29871 29941
|
||||||
|
29871 29941 29941
|
||||||
|
29871 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 29941 29941 29941 29941 29941 29941 29941 29941 29941
|
||||||
|
29871 13 29871 13 13 29871 13 13 13 29871 12 29871 12 12 29871 12 13 259 13 1678 13 268 13 418 13 243 162 157 131 313 8945 29897 29871 243 162 155 185 30722 243 162 143 174 30598 313 20787 953 3848 275 16125 630 29897 29871 31681 29871 243 162 169 156 243 162 169 156 29871 29941 29871 29941 29941 29871 29941 29941 29941 29871 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29941 29871 29941 29941 29941 29941 29941 29941 29941 29941 29871 29941 29889 29941 29871 29941 636 29941 29871 29941 856 29941 29871 31849 31324 31934 228 162 142 228 161 146 228 162 133 228 161 153 228 161 186 31708 228 162 132 31708 228 161 165 31324 228 161 136 243 162 155 132 1577 30672 31522 30505 11548 31041 30732 29896 29941 29896 29946 29896 29945 29896 30408 30739 448 23648 2751 25512 1538 4851 665 1386 29713 1305 14550 4907 11120 16159 16159 16159 15945 15945 3045 636 6824 6824 6824 8773 8773 8773 306 29915 345 1063 525 29873 1025 540 29915 29879 727 29892 525 1525 366 1854 29973 525 29924 451 1854 306 29915 645 1207 372 29892 525 29928 366 763 777 23429 29973 1334 29915 29963 29872 263 29915 29880 29931
|
||||||
Binary file not shown.
102
models/ggml-vocab-starcoder.gguf.inp
Normal file
102
models/ggml-vocab-starcoder.gguf.inp
Normal file
@ -0,0 +1,102 @@
|
|||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello world
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello World!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, world!
|
||||||
|
__ggml_vocab_test__
|
||||||
|
this is 🦙.cpp
|
||||||
|
__ggml_vocab_test__
|
||||||
|
w048 7tuijk dsdfhu
|
||||||
|
__ggml_vocab_test__
|
||||||
|
нещо на Български
|
||||||
|
__ggml_vocab_test__
|
||||||
|
កាន់តែពិសេសអាចខលចេញ
|
||||||
|
__ggml_vocab_test__
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello
|
||||||
|
Hello
|
||||||
|
__ggml_vocab_test__
|
||||||
|
(
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
=
|
||||||
|
__ggml_vocab_test__
|
||||||
|
' era
|
||||||
|
__ggml_vocab_test__
|
||||||
|
Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
3333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
33333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
333333333
|
||||||
|
__ggml_vocab_test__
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
|
||||||
|
🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
|
||||||
|
__ggml_vocab_test__
|
||||||
41
models/ggml-vocab-starcoder.gguf.out
Normal file
41
models/ggml-vocab-starcoder.gguf.out
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
|
||||||
|
244
|
||||||
|
280
|
||||||
|
283
|
||||||
|
221
|
||||||
|
222
|
||||||
|
499
|
||||||
|
3067
|
||||||
|
15767
|
||||||
|
8302 5810
|
||||||
|
12009 5810
|
||||||
|
8302 10914
|
||||||
|
12009 10914
|
||||||
|
12009 10914 38
|
||||||
|
8302 49 5810 38
|
||||||
|
12009 49 5810 38
|
||||||
|
477 458 5954 137 271 51 3779
|
||||||
|
124 53 57 61 244 60 121 1726 12568 10240 1519 8290
|
||||||
|
39916 8389 1059 9504 40216 13858 2073 8983 12571 1539 10721
|
||||||
|
14566 246 14566 152 14566 265 30428 257 14566 261 30428 248 14566 268 14566 153 14566 277 30428 247 14566 277 14566 133 14566 152 14566 251 14566 247 14566 273 14566 251 30428 247 14566 255
|
||||||
|
3822 272 246 327 4434 46 18445 152 46030 45022 142 13878 327 12585 19884 33773 40920 751 46 41839 327 2605 22716 708 1421 2840 4387 2421 46
|
||||||
|
8302
|
||||||
|
12009
|
||||||
|
244 12009
|
||||||
|
280 12009
|
||||||
|
283 12009
|
||||||
|
283 12009 303 12009
|
||||||
|
327
|
||||||
|
222 299
|
||||||
|
44 34719
|
||||||
|
8302 49 553 44 483 38 4998 904 863 18445 247 1037 4995 13379 2924 9515 17823 54 56 54 57 54 58 54 11904 47892
|
||||||
|
56
|
||||||
|
56 56
|
||||||
|
56 56 56
|
||||||
|
56 56 56 56
|
||||||
|
56 56 56 56 56
|
||||||
|
56 56 56 56 56 56
|
||||||
|
56 56 56 56 56 56 56
|
||||||
|
56 56 56 56 56 56 56 56
|
||||||
|
56 56 56 56 56 56 56 56 56
|
||||||
|
353 736 8886 221 10883 4238 16101 28540 222 3822 272 246 327 4434 46 18445 152 46030 45022 142 13878 327 12585 19884 33773 40920 751 46 41839 5954 137 271 3822 137 271 244 56 244 56 56 244 56 56 56 244 56 56 56 56 244 56 56 56 56 56 244 56 56 56 56 56 56 244 56 56 56 56 56 56 56 244 56 56 56 56 56 56 56 56 244 56 51 56 244 56 516 56 244 56 1198 56 244 14566 246 14566 152 14566 265 30428 257 14566 261 30428 248 14566 268 14566 153 14566 277 30428 247 14566 277 14566 133 14566 152 14566 251 36570 247 1037 4995 13379 2924 9515 17823 54 56 54 57 54 58 54 11904 47892 20895 16625 13047 8389 1059 9504 40216 13858 2073 8983 12571 1539 10721 5918 9643 13298 932 31723 31330 9221 3226 35426 10400 457 4783 2602 349 121 1477 957 1200 2038 49 349 632 863 3673 68 349 82 666 3673 457 4650 1949 580 49 349 73 863 2144 1649 35941 68 2726 44 7728 331 44 113 81
|
||||||
@ -7,6 +7,7 @@
|
|||||||
-r ./requirements/requirements-convert.txt
|
-r ./requirements/requirements-convert.txt
|
||||||
|
|
||||||
-r ./requirements/requirements-convert-hf-to-gguf.txt
|
-r ./requirements/requirements-convert-hf-to-gguf.txt
|
||||||
|
-r ./requirements/requirements-convert-hf-to-gguf-update.txt
|
||||||
-r ./requirements/requirements-convert-llama-ggml-to-gguf.txt
|
-r ./requirements/requirements-convert-llama-ggml-to-gguf.txt
|
||||||
-r ./requirements/requirements-convert-lora-to-ggml.txt
|
-r ./requirements/requirements-convert-lora-to-ggml.txt
|
||||||
-r ./requirements/requirements-convert-persimmon-to-gguf.txt
|
-r ./requirements/requirements-convert-persimmon-to-gguf.txt
|
||||||
|
|||||||
3
requirements/requirements-convert-hf-to-gguf-update.txt
Normal file
3
requirements/requirements-convert-hf-to-gguf-update.txt
Normal file
@ -0,0 +1,3 @@
|
|||||||
|
-r ./requirements-convert.txt
|
||||||
|
torch~=2.1.1
|
||||||
|
einops~=0.7.0
|
||||||
@ -168,6 +168,11 @@ fi
|
|||||||
|
|
||||||
check_convert_script convert.py
|
check_convert_script convert.py
|
||||||
for py in convert-*.py; do
|
for py in convert-*.py; do
|
||||||
|
# skip convert-hf-to-gguf-update.py
|
||||||
|
# TODO: the check is failing for some reason:
|
||||||
|
# https://github.com/ggerganov/llama.cpp/actions/runs/8875330981/job/24364557177?pr=6920
|
||||||
|
[[ $py == convert-hf-to-gguf-update.py ]] && continue
|
||||||
|
|
||||||
check_convert_script "$py"
|
check_convert_script "$py"
|
||||||
done
|
done
|
||||||
|
|
||||||
|
|||||||
@ -1,10 +1,40 @@
|
|||||||
|
function(llama_test target)
|
||||||
|
include(CMakeParseArguments)
|
||||||
|
set(options)
|
||||||
|
set(oneValueArgs NAME LABEL WORKING_DIRECTORY)
|
||||||
|
set(multiValueArgs ARGS)
|
||||||
|
cmake_parse_arguments(LLAMA_TEST "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
|
||||||
|
|
||||||
|
if (NOT DEFINED LLAMA_TEST_LABEL)
|
||||||
|
set(LLAMA_TEST_LABEL "main")
|
||||||
|
endif()
|
||||||
|
if (NOT DEFINED LLAMA_TEST_WORKING_DIRECTORY)
|
||||||
|
set(LLAMA_TEST_WORKING_DIRECTORY .)
|
||||||
|
endif()
|
||||||
|
if (DEFINED LLAMA_TEST_NAME)
|
||||||
|
set(TEST_NAME ${LLAMA_TEST_NAME})
|
||||||
|
else()
|
||||||
|
set(TEST_NAME ${target})
|
||||||
|
endif()
|
||||||
|
|
||||||
|
set(TEST_TARGET ${target})
|
||||||
|
|
||||||
|
add_test(
|
||||||
|
NAME ${TEST_NAME}
|
||||||
|
WORKING_DIRECTORY ${LLAMA_TEST_WORKING_DIRECTORY}
|
||||||
|
COMMAND $<TARGET_FILE:${TEST_TARGET}>
|
||||||
|
${LLAMA_TEST_ARGS})
|
||||||
|
|
||||||
|
set_property(TEST ${TEST_NAME} PROPERTY LABELS ${LLAMA_TEST_LABEL})
|
||||||
|
endfunction()
|
||||||
|
|
||||||
# Builds and runs a test source file.
|
# Builds and runs a test source file.
|
||||||
# Optional args:
|
# Optional args:
|
||||||
# - NAME: name of the executable & test target (defaults to the source file name without extension)
|
# - NAME: name of the executable & test target (defaults to the source file name without extension)
|
||||||
# - LABEL: label for the test (defaults to main)
|
# - LABEL: label for the test (defaults to main)
|
||||||
# - ARGS: arguments to pass to the test executable
|
# - ARGS: arguments to pass to the test executable
|
||||||
# - WORKING_DIRECTORY
|
# - WORKING_DIRECTORY
|
||||||
function(llama_test source)
|
function(llama_target_and_test source)
|
||||||
include(CMakeParseArguments)
|
include(CMakeParseArguments)
|
||||||
set(options)
|
set(options)
|
||||||
set(oneValueArgs NAME LABEL WORKING_DIRECTORY)
|
set(oneValueArgs NAME LABEL WORKING_DIRECTORY)
|
||||||
@ -35,41 +65,67 @@ function(llama_test source)
|
|||||||
set_property(TEST ${TEST_TARGET} PROPERTY LABELS ${LLAMA_TEST_LABEL})
|
set_property(TEST ${TEST_TARGET} PROPERTY LABELS ${LLAMA_TEST_LABEL})
|
||||||
endfunction()
|
endfunction()
|
||||||
|
|
||||||
# llama_test(test-double-float.cpp) # SLOW
|
# build test-tokenizer-0 target once and add many tests
|
||||||
llama_test(test-quantize-fns.cpp)
|
add_executable(test-tokenizer-0 test-tokenizer-0.cpp)
|
||||||
llama_test(test-quantize-perf.cpp)
|
target_link_libraries(test-tokenizer-0 PRIVATE common)
|
||||||
llama_test(test-sampling.cpp)
|
install(TARGETS test-tokenizer-0 RUNTIME)
|
||||||
llama_test(test-chat-template.cpp)
|
|
||||||
|
|
||||||
llama_test(test-tokenizer-0-llama.cpp NAME test-tokenizer-0-llama ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama.gguf)
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-llama-spm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-spm.gguf)
|
||||||
llama_test(test-tokenizer-0-falcon.cpp NAME test-tokenizer-0-falcon ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-llama-bpe ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-bpe.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-phi-3 ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-phi-3.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-falcon ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-deepseek-llm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-deepseek-llm.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-deepseek-coder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-deepseek-coder.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-bert-bge ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-bert-bge.gguf)
|
||||||
|
# TODO: enable when fixed
|
||||||
|
#llama_test(test-tokenizer-0 NAME test-tokenizer-0-mpt ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-mpt.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-starcoder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
|
||||||
|
llama_test(test-tokenizer-0 NAME test-tokenizer-0-gpt-2 ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-2.gguf)
|
||||||
|
|
||||||
llama_test(test-tokenizer-1-llama.cpp NAME test-tokenizer-1-llama ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama.gguf)
|
# build test-tokenizer-1-bpe target once and add many tests
|
||||||
llama_test(test-tokenizer-1-llama.cpp NAME test-tokenizer-1-baichuan ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-baichuan.gguf)
|
add_executable(test-tokenizer-1-bpe test-tokenizer-1-bpe.cpp)
|
||||||
|
target_link_libraries(test-tokenizer-1-bpe PRIVATE common)
|
||||||
|
install(TARGETS test-tokenizer-1-bpe RUNTIME)
|
||||||
|
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-falcon ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
|
# TODO: disabled due to slowness
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-aquila ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-aquila.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-llama-bpe ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-bpe.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-mpt ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-mpt.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-falcon ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-falcon.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-stablelm-3b-4e1t ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-stablelm-3b-4e1t.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-aquila ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-aquila.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-gpt-neox ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-neox.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-mpt ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-mpt.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-refact ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-refact.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-stablelm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-stablelm.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-starcoder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt-neox ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt-neox.gguf)
|
||||||
llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-gpt2 ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt2.gguf)
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-refact ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-refact.gguf)
|
||||||
#llama_test(test-tokenizer-1-bpe.cpp NAME test-tokenizer-1-bloom ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-bloom.gguf) # BIG
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-starcoder ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-starcoder.gguf)
|
||||||
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-gpt2 ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-gpt2.gguf)
|
||||||
|
#llama_test(test-tokenizer-1-bpe NAME test-tokenizer-1-bloom ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-bloom.gguf)
|
||||||
|
|
||||||
llama_test(test-grammar-parser.cpp)
|
# build test-tokenizer-1-spm target once and add many tests
|
||||||
llama_test(test-llama-grammar.cpp)
|
add_executable(test-tokenizer-1-spm test-tokenizer-1-spm.cpp)
|
||||||
llama_test(test-grammar-integration.cpp)
|
target_link_libraries(test-tokenizer-1-spm PRIVATE common)
|
||||||
llama_test(test-grad0.cpp)
|
install(TARGETS test-tokenizer-1-spm RUNTIME)
|
||||||
# llama_test(test-opt.cpp) # SLOW
|
|
||||||
llama_test(test-backend-ops.cpp)
|
|
||||||
|
|
||||||
llama_test(test-rope.cpp)
|
llama_test(test-tokenizer-1-spm NAME test-tokenizer-1-llama-spm ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-llama-spm.gguf)
|
||||||
|
#llama_test(test-tokenizer-1-spm NAME test-tokenizer-1-baichuan ARGS ${CMAKE_CURRENT_SOURCE_DIR}/../models/ggml-vocab-baichuan.gguf)
|
||||||
|
|
||||||
llama_test(test-model-load-cancel.cpp LABEL "model")
|
# llama_target_and_test(test-double-float.cpp) # SLOW
|
||||||
llama_test(test-autorelease.cpp LABEL "model")
|
llama_target_and_test(test-quantize-fns.cpp)
|
||||||
|
llama_target_and_test(test-quantize-perf.cpp)
|
||||||
|
llama_target_and_test(test-sampling.cpp)
|
||||||
|
llama_target_and_test(test-chat-template.cpp)
|
||||||
|
|
||||||
llama_test(test-json-schema-to-grammar.cpp WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/..)
|
llama_target_and_test(test-grammar-parser.cpp)
|
||||||
|
llama_target_and_test(test-llama-grammar.cpp)
|
||||||
|
llama_target_and_test(test-grammar-integration.cpp)
|
||||||
|
llama_target_and_test(test-grad0.cpp)
|
||||||
|
# llama_target_and_test(test-opt.cpp) # SLOW
|
||||||
|
llama_target_and_test(test-backend-ops.cpp)
|
||||||
|
|
||||||
|
llama_target_and_test(test-rope.cpp)
|
||||||
|
|
||||||
|
llama_target_and_test(test-model-load-cancel.cpp LABEL "model")
|
||||||
|
llama_target_and_test(test-autorelease.cpp LABEL "model")
|
||||||
|
|
||||||
|
llama_target_and_test(test-json-schema-to-grammar.cpp WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/..)
|
||||||
target_include_directories(test-json-schema-to-grammar PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/../examples/server)
|
target_include_directories(test-json-schema-to-grammar PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/../examples/server)
|
||||||
|
|
||||||
# dummy executable - not installed
|
# dummy executable - not installed
|
||||||
|
|||||||
@ -1,4 +1,11 @@
|
|||||||
# tests with BPE tokenizer
|
# tests with BPE tokenizer
|
||||||
|
#
|
||||||
|
# sample usage:
|
||||||
|
#
|
||||||
|
# python3 tests/test-tokenizer-0-bpe.py ~/Data/huggingface/Meta-Llama-3-8B-Instruct/
|
||||||
|
# python3 tests/test-tokenizer-0-bpe.py ~/Data/huggingface/falcon-7b/
|
||||||
|
# python3 tests/test-tokenizer-0-bpe.py ~/Data/huggingface/deepseek-coder-6.7b-instruct/
|
||||||
|
#
|
||||||
|
|
||||||
import argparse
|
import argparse
|
||||||
|
|
||||||
@ -20,6 +27,8 @@ tests = [
|
|||||||
" ",
|
" ",
|
||||||
"\t",
|
"\t",
|
||||||
"\n",
|
"\n",
|
||||||
|
"\n\n",
|
||||||
|
"\n\n\n",
|
||||||
"\t\n",
|
"\t\n",
|
||||||
"Hello world",
|
"Hello world",
|
||||||
" Hello world",
|
" Hello world",
|
||||||
@ -39,8 +48,19 @@ tests = [
|
|||||||
" Hello",
|
" Hello",
|
||||||
" Hello",
|
" Hello",
|
||||||
" Hello\n Hello",
|
" Hello\n Hello",
|
||||||
|
" (",
|
||||||
"\n =",
|
"\n =",
|
||||||
"' era",
|
"' era",
|
||||||
|
"Hello, y'all! How are you 😁 ?我想在apple工作1314151天～",
|
||||||
|
"3",
|
||||||
|
"33",
|
||||||
|
"333",
|
||||||
|
"3333",
|
||||||
|
"33333",
|
||||||
|
"333333",
|
||||||
|
"3333333",
|
||||||
|
"33333333",
|
||||||
|
"333333333",
|
||||||
]
|
]
|
||||||
|
|
||||||
for text in tests:
|
for text in tests:
|
||||||
@ -76,7 +96,22 @@ if fname_tok:
|
|||||||
# write to file
|
# write to file
|
||||||
with open(fname_out, 'w', encoding='utf-8') as f:
|
with open(fname_out, 'w', encoding='utf-8') as f:
|
||||||
for x in res:
|
for x in res:
|
||||||
f.write(str(x) + ' \'' + tokenizer.decode(x) + '\'\n')
|
# LLaMA v3 for some reason strips the space for these tokens (and others)
|
||||||
|
# if x == 662:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# elif x == 1174:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# elif x == 2564:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# elif x == 758:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# elif x == 949:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# elif x == 5354:
|
||||||
|
# f.write(str(x) + ' \' ' + tokenizer.decode(x) + '\'\n')
|
||||||
|
# else:
|
||||||
|
# f.write(str(x) + ' \'' + tokenizer.decode(x) + '\'\n')
|
||||||
|
f.write(str(x) + ' \'' + tokenizer.decode(x).strip() + '\'\n')
|
||||||
print('len(res): ', len(res))
|
print('len(res): ', len(res))
|
||||||
print('len(lines): ', len(lines))
|
print('len(lines): ', len(lines))
|
||||||
print('results written to: ', fname_out)
|
print('results written to: ', fname_out)
|
||||||
@ -1,187 +0,0 @@
|
|||||||
#include "llama.h"
|
|
||||||
#include "common.h"
|
|
||||||
#include "console.h"
|
|
||||||
|
|
||||||
#include <cstdio>
|
|
||||||
#include <string>
|
|
||||||
#include <map>
|
|
||||||
#include <vector>
|
|
||||||
#include <fstream>
|
|
||||||
|
|
||||||
// generate using test-tokenizer-0-falcon.py
|
|
||||||
static const std::map<std::string, std::vector<llama_token>> & k_tests() {
|
|
||||||
static std::map<std::string, std::vector<llama_token>> _k_tests = {
|
|
||||||
{ "" , { }, },
|
|
||||||
{ " " , { 204, }, },
|
|
||||||
{ " " , { 258, }, },
|
|
||||||
{ " " , { 466, }, },
|
|
||||||
{ "\t" , { 192, }, },
|
|
||||||
{ "\n" , { 193, }, },
|
|
||||||
{ "\t\n" , { 19125, }, },
|
|
||||||
{ "Hello world" , { 9856, 1079, }, },
|
|
||||||
{ " Hello world" , { 23090, 1079, }, },
|
|
||||||
{ "Hello World" , { 9856, 2889, }, },
|
|
||||||
{ " Hello World" , { 23090, 2889, }, },
|
|
||||||
{ " Hello World!" , { 23090, 2889, 12, }, },
|
|
||||||
{ "Hello, world!" , { 9856, 23, 1079, 12, }, },
|
|
||||||
{ " Hello, world!" , { 23090, 23, 1079, 12, }, },
|
|
||||||
{ " this is 🦙.cpp" , { 414, 304, 3346, 111, 231, 25, 29247, }, },
|
|
||||||
{ "w048 7tuijk dsdfhu" , { 98, 55866, 204, 34, 16682, 7149, 36190, 6869, 11481, }, },
|
|
||||||
{ "нещо на Български" , { 150, 133, 6207, 151, 215, 150, 134, 5052, 133, 6279, 5052, 223, 151, 216, 49679, 123, 53110, 47043, 7795, }, },
|
|
||||||
{ "កាន់តែពិសេសអាចខលចេញ" , { 38154, 206, 38154, 126, 38154, 225, 167, 237, 217, 38154, 221, 167, 237, 208, 38154, 228, 38154, 127, 38154, 237, 167, 237, 207, 38154, 237, 38154, 107, 38154, 126, 38154, 211, 38154, 207, 38154, 233, 38154, 211, 167, 237, 207, 38154, 215, }, },
|
|
||||||
{ "🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)", { 2571, 232, 206, 204, 19, 11003, 20, 8196, 126, 283, 219, 48778, 116, 13392, 204, 19, 51831, 732, 63209, 1741, 7955, 522, 20, 22438, 211, 204, 19, 7927, 53360, 325, 504, 701, 946, 10930, 20, }, },
|
|
||||||
{ "Hello" , { 9856, }, },
|
|
||||||
{ " Hello" , { 23090, }, },
|
|
||||||
{ " Hello" , { 204, 23090, }, },
|
|
||||||
{ " Hello" , { 258, 23090, }, },
|
|
||||||
{ " Hello" , { 466, 23090, }, },
|
|
||||||
{ " Hello\n Hello" , { 466, 23090, 742, 23090, }, },
|
|
||||||
{ "\n =" , { 1212, 40, }, },
|
|
||||||
{ "' era" , { 18, 4932, }, },
|
|
||||||
};
|
|
||||||
|
|
||||||
return _k_tests;
|
|
||||||
}
|
|
||||||
|
|
||||||
int main(int argc, char **argv) {
|
|
||||||
if (argc < 2) {
|
|
||||||
fprintf(stderr, "Usage: %s vocab-file [text-file]\n", argv[0]);
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
const std::string fname = argv[1];
|
|
||||||
|
|
||||||
std::string fname_text;
|
|
||||||
if (argc > 2) {
|
|
||||||
fname_text = argv[2];
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : reading vocab from: '%s'\n", __func__, fname.c_str());
|
|
||||||
|
|
||||||
llama_model * model;
|
|
||||||
llama_context * ctx;
|
|
||||||
|
|
||||||
llama_backend_init();
|
|
||||||
|
|
||||||
// load the vocab
|
|
||||||
{
|
|
||||||
auto mparams = llama_model_default_params();
|
|
||||||
|
|
||||||
mparams.vocab_only = true;
|
|
||||||
|
|
||||||
model = llama_load_model_from_file(fname.c_str(), mparams);
|
|
||||||
|
|
||||||
if (model == NULL) {
|
|
||||||
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
auto cparams = llama_context_default_params();
|
|
||||||
|
|
||||||
ctx = llama_new_context_with_model(model, cparams);
|
|
||||||
|
|
||||||
if (ctx == NULL) {
|
|
||||||
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
|
||||||
llama_free_model(model);
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_BPE) {
|
|
||||||
fprintf(stderr, "%s : error: vocab type is not BPE\n", __func__);
|
|
||||||
llama_free_model(model);
|
|
||||||
llama_free(ctx);
|
|
||||||
return 2;
|
|
||||||
}
|
|
||||||
|
|
||||||
#ifdef _WIN32
|
|
||||||
// We need this for unicode console support
|
|
||||||
console::init(false, false);
|
|
||||||
atexit([]() { console::cleanup(); });
|
|
||||||
#endif
|
|
||||||
|
|
||||||
bool success = true;
|
|
||||||
|
|
||||||
for (const auto & test_kv : k_tests()) {
|
|
||||||
const std::vector<llama_token> res = llama_tokenize(ctx, test_kv.first, false);
|
|
||||||
|
|
||||||
printf("\n");
|
|
||||||
printf("src: '%s'\n", test_kv.first.c_str());
|
|
||||||
printf("res: '%s'\n", llama_detokenize_bpe(ctx, res).c_str());
|
|
||||||
printf("tok: ");
|
|
||||||
for (const auto & tok : res) {
|
|
||||||
printf("%d ", tok);
|
|
||||||
}
|
|
||||||
printf("\n");
|
|
||||||
|
|
||||||
bool correct = res.size() == test_kv.second.size();
|
|
||||||
|
|
||||||
for (int i = 0; i < (int) res.size() && correct; ++i) {
|
|
||||||
if (test_kv.second[i] != res[i]) {
|
|
||||||
correct = false;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!correct) {
|
|
||||||
fprintf(stderr, "%s : failed test: '%s'\n", __func__, test_kv.first.c_str());
|
|
||||||
fprintf(stderr, "%s : detokenized to: '%s' instead of '%s'\n", __func__,
|
|
||||||
llama_detokenize_bpe(ctx, res).c_str(),
|
|
||||||
llama_detokenize_bpe(ctx, test_kv.second).c_str());
|
|
||||||
fprintf(stderr, "%s : expected tokens: ", __func__);
|
|
||||||
for (const auto & t : test_kv.second) {
|
|
||||||
fprintf(stderr, "%6d, ", t);
|
|
||||||
}
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
fprintf(stderr, "%s : got tokens: ", __func__);
|
|
||||||
for (const auto & t : res) {
|
|
||||||
fprintf(stderr, "%6d, ", t);
|
|
||||||
}
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
|
|
||||||
success = false;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!fname_text.empty()) {
|
|
||||||
fprintf(stderr, "%s : tokenizing: '%s'\n", __func__, fname_text.c_str());
|
|
||||||
|
|
||||||
std::string text;
|
|
||||||
{
|
|
||||||
std::ifstream ifs(fname_text);
|
|
||||||
if (!ifs) {
|
|
||||||
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_text.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
text = std::string(std::istreambuf_iterator<char>(ifs), std::istreambuf_iterator<char>());
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : text size: %zu\n", __func__, text.size());
|
|
||||||
|
|
||||||
const std::vector<llama_token> res = llama_tokenize(ctx, text, false);
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : tokens: %zu\n", __func__, res.size());
|
|
||||||
|
|
||||||
{
|
|
||||||
const std::string fname_out = fname_text + ".tokcpp";
|
|
||||||
|
|
||||||
std::ofstream ofs(fname_out);
|
|
||||||
if (!ofs) {
|
|
||||||
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_out.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
for (const auto & tok : res) {
|
|
||||||
ofs << tok << " '" << llama_detokenize_bpe(ctx, std::vector<int>{tok}) << "'" << std::endl;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());
|
|
||||||
}
|
|
||||||
|
|
||||||
llama_free_model(model);
|
|
||||||
llama_free(ctx);
|
|
||||||
|
|
||||||
llama_backend_free();
|
|
||||||
|
|
||||||
return success ? 0 : 3;
|
|
||||||
}
|
|
||||||
@ -1,190 +0,0 @@
|
|||||||
#include "llama.h"
|
|
||||||
#include "common.h"
|
|
||||||
#include "console.h"
|
|
||||||
|
|
||||||
#include <cstdio>
|
|
||||||
#include <string>
|
|
||||||
#include <map>
|
|
||||||
#include <vector>
|
|
||||||
#include <fstream>
|
|
||||||
|
|
||||||
// generate using test-tokenizer-0-llama.py
|
|
||||||
static const std::map<std::string, std::vector<llama_token>> & k_tests() {
|
|
||||||
static std::map<std::string, std::vector<llama_token>> _k_tests = {
|
|
||||||
{ "" , { }, },
|
|
||||||
{ " " , { 259, }, },
|
|
||||||
{ " " , { 1678, }, },
|
|
||||||
{ " " , { 268, }, },
|
|
||||||
{ "\t" , { 29871, 12, }, },
|
|
||||||
{ "\n" , { 29871, 13, }, },
|
|
||||||
{ "\t\n" , { 29871, 12, 13, }, },
|
|
||||||
{ "Hello world" , { 15043, 3186, }, },
|
|
||||||
{ " Hello world" , { 29871, 15043, 3186, }, },
|
|
||||||
{ "Hello World" , { 15043, 2787, }, },
|
|
||||||
{ " Hello World" , { 29871, 15043, 2787, }, },
|
|
||||||
{ " Hello World!" , { 29871, 15043, 2787, 29991, }, },
|
|
||||||
{ "Hello, world!" , { 15043, 29892, 3186, 29991, }, },
|
|
||||||
{ " Hello, world!" , { 29871, 15043, 29892, 3186, 29991, }, },
|
|
||||||
{ " this is 🦙.cpp" , { 29871, 445, 338, 29871, 243, 162, 169, 156, 29889, 8223, }, },
|
|
||||||
{ "w048 7tuijk dsdfhu" , { 281, 29900, 29946, 29947, 29871, 29955, 9161, 13535, 18031, 2176, 6905, }, },
|
|
||||||
{ "нещо на Български" , { 1538, 4851, 665, 1386, 29713, 1305, }, },
|
|
||||||
{ "កាន់តែពិសេសអាចខលចេញ" , { 29871, 31849, 31324, 31934, 228, 162, 142, 228, 161, 146, 228, 162, 133, 228, 161, 153, 228, 161, 186, 31708, 228, 162, 132, 31708, 228, 161, 165, 31324, 228, 161, 136, 228, 161, 132, 228, 161, 158, 228, 161, 136, 228, 162, 132, 228, 161, 140, }, },
|
|
||||||
{ "🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)", { 29871, 243, 162, 157, 131, 313, 8945, 29897, 29871, 243, 162, 155, 185, 30722, 243, 162, 143, 174, 30598, 313, 20787, 953, 3848, 275, 16125, 630, 29897, 29871, 31681, 313, 6194, 953, 29877, 2397, 393, 756, 967, 1914, 5993, 29897, }, },
|
|
||||||
{ "Hello" , { 15043, }, },
|
|
||||||
{ " Hello" , { 29871, 15043, }, },
|
|
||||||
{ " Hello" , { 259, 15043, }, },
|
|
||||||
{ " Hello" , { 1678, 15043, }, },
|
|
||||||
{ " Hello" , { 268, 15043, }, },
|
|
||||||
{ " Hello\n Hello" , { 268, 15043, 13, 1678, 15043, }, },
|
|
||||||
{ " (" , { 29871, 313, }, },
|
|
||||||
};
|
|
||||||
|
|
||||||
return _k_tests;
|
|
||||||
}
|
|
||||||
|
|
||||||
int main(int argc, char **argv) {
|
|
||||||
if (argc < 2) {
|
|
||||||
fprintf(stderr, "Usage: %s vocab-file [text-file]\n", argv[0]);
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
const std::string fname = argv[1];
|
|
||||||
|
|
||||||
std::string fname_text;
|
|
||||||
if (argc > 2) {
|
|
||||||
fname_text = argv[2];
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : reading vocab from: '%s'\n", __func__, fname.c_str());
|
|
||||||
|
|
||||||
llama_model * model;
|
|
||||||
llama_context * ctx;
|
|
||||||
|
|
||||||
llama_backend_init();
|
|
||||||
|
|
||||||
// load the vocab
|
|
||||||
{
|
|
||||||
auto mparams = llama_model_default_params();
|
|
||||||
|
|
||||||
mparams.vocab_only = true;
|
|
||||||
|
|
||||||
model = llama_load_model_from_file(fname.c_str(), mparams);
|
|
||||||
|
|
||||||
if (model == NULL) {
|
|
||||||
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
auto cparams = llama_context_default_params();
|
|
||||||
|
|
||||||
ctx = llama_new_context_with_model(model, cparams);
|
|
||||||
|
|
||||||
if (ctx == NULL) {
|
|
||||||
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
|
||||||
llama_free_model(model);
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (llama_vocab_type(model) != LLAMA_VOCAB_TYPE_SPM) {
|
|
||||||
fprintf(stderr, "%s : error: vocab type is not SPM\n", __func__);
|
|
||||||
llama_free_model(model);
|
|
||||||
llama_free(ctx);
|
|
||||||
return 2;
|
|
||||||
}
|
|
||||||
|
|
||||||
#ifdef _WIN32
|
|
||||||
// We need this for unicode console support
|
|
||||||
console::init(false, false);
|
|
||||||
atexit([]() { console::cleanup(); });
|
|
||||||
#endif
|
|
||||||
|
|
||||||
bool success = true;
|
|
||||||
|
|
||||||
for (const auto & test_kv : k_tests()) {
|
|
||||||
const std::vector<llama_token> res_bos = llama_tokenize(ctx, test_kv.first, true);
|
|
||||||
const std::vector<llama_token> res_nobos = llama_tokenize(ctx, test_kv.first, false);
|
|
||||||
|
|
||||||
printf("\n");
|
|
||||||
printf("src: '%s'\n", test_kv.first.c_str());
|
|
||||||
printf("res: '%s'\n", llama_detokenize_spm(ctx, res_bos).c_str());
|
|
||||||
printf("tok: ");
|
|
||||||
for (const auto & tok : res_bos) {
|
|
||||||
printf("%d ", tok);
|
|
||||||
}
|
|
||||||
printf("\n");
|
|
||||||
|
|
||||||
bool correct = res_nobos.size() == test_kv.second.size() && res_bos.size() == res_nobos.size() + 1 && res_bos[0] == 1;
|
|
||||||
|
|
||||||
for (int i = 0; i < (int) res_nobos.size() && correct; ++i) {
|
|
||||||
if (test_kv.second[i] != res_bos[i + 1]) {
|
|
||||||
correct = false;
|
|
||||||
}
|
|
||||||
if (test_kv.second[i] != res_nobos[i]) {
|
|
||||||
correct = false;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!correct) {
|
|
||||||
fprintf(stderr, "%s : failed test: '%s'\n", __func__, test_kv.first.c_str());
|
|
||||||
fprintf(stderr, "%s : detokenized to: '%s' instead of '%s'\n", __func__,
|
|
||||||
llama_detokenize_spm(ctx, res_nobos).c_str(),
|
|
||||||
llama_detokenize_spm(ctx, test_kv.second).c_str());
|
|
||||||
fprintf(stderr, "%s : expected tokens: ", __func__);
|
|
||||||
for (const auto & t : test_kv.second) {
|
|
||||||
fprintf(stderr, "%6d, ", t);
|
|
||||||
}
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
fprintf(stderr, "%s : got tokens: ", __func__);
|
|
||||||
for (const auto & t : res_nobos) {
|
|
||||||
fprintf(stderr, "%6d, ", t);
|
|
||||||
}
|
|
||||||
fprintf(stderr, "\n");
|
|
||||||
|
|
||||||
success = false;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
if (!fname_text.empty()) {
|
|
||||||
fprintf(stderr, "%s : tokenizing: '%s'\n", __func__, fname_text.c_str());
|
|
||||||
|
|
||||||
std::string text;
|
|
||||||
{
|
|
||||||
std::ifstream ifs(fname_text);
|
|
||||||
if (!ifs) {
|
|
||||||
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_text.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
text = std::string(std::istreambuf_iterator<char>(ifs), std::istreambuf_iterator<char>());
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : text size: %zu\n", __func__, text.size());
|
|
||||||
|
|
||||||
const std::vector<llama_token> res = llama_tokenize(ctx, text, true);
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : tokens: %zu\n", __func__, res.size());
|
|
||||||
|
|
||||||
{
|
|
||||||
const std::string fname_out = fname_text + ".tokcpp";
|
|
||||||
|
|
||||||
std::ofstream ofs(fname_out);
|
|
||||||
if (!ofs) {
|
|
||||||
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_out.c_str());
|
|
||||||
return 1;
|
|
||||||
}
|
|
||||||
|
|
||||||
for (const auto & tok : res) {
|
|
||||||
ofs << tok << " '" << llama_detokenize_spm(ctx, std::vector<int>{tok}) << "'" << std::endl;
|
|
||||||
}
|
|
||||||
}
|
|
||||||
|
|
||||||
fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());
|
|
||||||
}
|
|
||||||
|
|
||||||
llama_free_model(model);
|
|
||||||
llama_free(ctx);
|
|
||||||
|
|
||||||
llama_backend_free();
|
|
||||||
|
|
||||||
return success ? 0 : 3;
|
|
||||||
}
|
|
||||||
@ -1,4 +1,11 @@
|
|||||||
# tests with SPM tokenizer
|
# tests with SPM tokenizer
|
||||||
|
#
|
||||||
|
# sample usage:
|
||||||
|
#
|
||||||
|
# python3 tests/test-tokenizer-0-spm.py ~/Data/huggingface/Llama-2-7b-hf/
|
||||||
|
# python3 tests/test-tokenizer-0-spm.py ~/Data/huggingface/CodeLlama-34b-Instruct-hf/
|
||||||
|
#
|
||||||
|
|
||||||
|
|
||||||
import argparse
|
import argparse
|
||||||
|
|
||||||
@ -20,6 +27,8 @@ tests = [
|
|||||||
" ",
|
" ",
|
||||||
"\t",
|
"\t",
|
||||||
"\n",
|
"\n",
|
||||||
|
"\n\n",
|
||||||
|
"\n\n\n",
|
||||||
"\t\n",
|
"\t\n",
|
||||||
"Hello world",
|
"Hello world",
|
||||||
" Hello world",
|
" Hello world",
|
||||||
@ -39,6 +48,19 @@ tests = [
|
|||||||
" Hello",
|
" Hello",
|
||||||
" Hello",
|
" Hello",
|
||||||
" Hello\n Hello",
|
" Hello\n Hello",
|
||||||
|
" (",
|
||||||
|
"\n =",
|
||||||
|
"' era",
|
||||||
|
"Hello, y'all! How are you 😁 ?我想在apple工作1314151天～",
|
||||||
|
"3",
|
||||||
|
"33",
|
||||||
|
"333",
|
||||||
|
"3333",
|
||||||
|
"33333",
|
||||||
|
"333333",
|
||||||
|
"3333333",
|
||||||
|
"33333333",
|
||||||
|
"333333333",
|
||||||
]
|
]
|
||||||
|
|
||||||
|
|
||||||
271
tests/test-tokenizer-0.cpp
Normal file
271
tests/test-tokenizer-0.cpp
Normal file
@ -0,0 +1,271 @@
|
|||||||
|
#include "llama.h"
|
||||||
|
#include "common.h"
|
||||||
|
#include "console.h"
|
||||||
|
|
||||||
|
#include <cstdio>
|
||||||
|
#include <string>
|
||||||
|
#include <map>
|
||||||
|
#include <vector>
|
||||||
|
#include <fstream>
|
||||||
|
|
||||||
|
//static const std::map<std::string, std::vector<llama_token>> & k_tests() {
|
||||||
|
// static std::map<std::string, std::vector<llama_token>> _k_tests = {
|
||||||
|
// { "" , { }, },
|
||||||
|
// { " " , { 220, }, },
|
||||||
|
// { " " , { 256, }, },
|
||||||
|
// { " " , { 262, }, },
|
||||||
|
// { "\t" , { 197, }, },
|
||||||
|
// { "\n" , { 198, }, },
|
||||||
|
// { "\n\n" , { 271, }, },
|
||||||
|
// { "\n\n\n" , { 1432, }, },
|
||||||
|
// { "\t\n" , { 1602, }, },
|
||||||
|
// { "Hello world" , { 9906, 1917, }, },
|
||||||
|
// { " Hello world" , { 22691, 1917, }, },
|
||||||
|
// { "Hello World" , { 9906, 4435, }, },
|
||||||
|
// { " Hello World" , { 22691, 4435, }, },
|
||||||
|
// { " Hello World!" , { 22691, 4435, 0, }, },
|
||||||
|
// { "Hello, world!" , { 9906, 11, 1917, 0, }, },
|
||||||
|
// { " Hello, world!" , { 22691, 11, 1917, 0, }, },
|
||||||
|
// { " this is 🦙.cpp" , { 420, 374, 11410, 99, 247, 13, 11055, }, },
|
||||||
|
// { "w048 7tuijk dsdfhu" , { 86, 23904, 220, 22, 83, 2005, 42908, 11729, 3013, 17156, }, },
|
||||||
|
// { "нещо на Български" , { 79862, 102118, 13373, 64571, 34694, 3114, 112203, 80112, }, },
|
||||||
|
// { "កាន់តែពិសេសអាចខលចេញ" , { 21549, 222, 98629, 241, 45358, 233, 21549, 237, 45358, 224, 21549, 244, 21549, 115, 21549, 253, 45358, 223, 21549, 253, 21549, 95, 98629, 227, 21549, 223, 21549, 249, 21549, 227, 45358, 223, 21549, 231, }, },
|
||||||
|
// { "🚀 (normal) 😶🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)", { 9468, 248, 222, 320, 8416, 8, 27623, 114, 102470, 9468, 234, 104, 31643, 320, 36773, 100166, 98634, 8, 26602, 227, 320, 3323, 43465, 430, 706, 1202, 1866, 4037, 8, }, },
|
||||||
|
// { "Hello" , { 9906, }, },
|
||||||
|
// { " Hello" , { 22691, }, },
|
||||||
|
// { " Hello" , { 220, 22691, }, },
|
||||||
|
// { " Hello" , { 256, 22691, }, },
|
||||||
|
// { " Hello" , { 262, 22691, }, },
|
||||||
|
// { " Hello\n Hello" , { 262, 22691, 198, 262, 22691, }, },
|
||||||
|
// { " (" , { 320, }, },
|
||||||
|
// { "\n =" , { 198, 284, }, },
|
||||||
|
// { "' era" , { 6, 11639, }, },
|
||||||
|
// { "Hello, y'all! How are you 😁 ?我想在apple工作1314151天～", { 9906, 11, 379, 65948, 0, 2650, 527, 499, 27623, 223, 949, 37046, 101067, 19000, 23182, 102301, 9263, 18136, 16, 36827, 21909, }, },
|
||||||
|
// { "3" , { 18, }, },
|
||||||
|
// { "33" , { 1644, }, },
|
||||||
|
// { "333" , { 8765, }, },
|
||||||
|
// { "3333" , { 8765, 18, }, },
|
||||||
|
// { "33333" , { 8765, 1644, }, },
|
||||||
|
// { "333333" , { 8765, 8765, }, },
|
||||||
|
// { "3333333" , { 8765, 8765, 18, }, },
|
||||||
|
// { "33333333" , { 8765, 8765, 1644, }, },
|
||||||
|
// { "333333333" , { 8765, 8765, 8765, }, },
|
||||||
|
// };
|
||||||
|
//
|
||||||
|
// return _k_tests;
|
||||||
|
//}
|
||||||
|
|
||||||
|
static std::map<std::string, std::vector<llama_token>> read_tests(const std::string & fname_inp, const std::string & fname_out) {
|
||||||
|
std::map<std::string, std::vector<llama_token>> tests;
|
||||||
|
|
||||||
|
std::ifstream ifs_inp(fname_inp);
|
||||||
|
if (!ifs_inp) {
|
||||||
|
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_inp.c_str());
|
||||||
|
return tests;
|
||||||
|
}
|
||||||
|
|
||||||
|
std::string sraw((std::istreambuf_iterator<char>(ifs_inp)), std::istreambuf_iterator<char>());
|
||||||
|
|
||||||
|
std::ifstream ifs_out(fname_out);
|
||||||
|
if (!ifs_out) {
|
||||||
|
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_out.c_str());
|
||||||
|
return tests;
|
||||||
|
}
|
||||||
|
|
||||||
|
std::vector<std::string> sout;
|
||||||
|
for (std::string line; std::getline(ifs_out, line);) {
|
||||||
|
sout.push_back(line);
|
||||||
|
}
|
||||||
|
|
||||||
|
const std::string sep = "\n__ggml_vocab_test__\n";
|
||||||
|
|
||||||
|
std::vector<std::string> sinp;
|
||||||
|
|
||||||
|
size_t pos = 0;
|
||||||
|
while (pos < sraw.size()) {
|
||||||
|
const size_t next = sraw.find(sep, pos);
|
||||||
|
if (next == std::string::npos) {
|
||||||
|
sinp.push_back(sraw.substr(pos));
|
||||||
|
break;
|
||||||
|
}
|
||||||
|
sinp.push_back(sraw.substr(pos, next - pos));
|
||||||
|
pos = next + sep.size();
|
||||||
|
}
|
||||||
|
|
||||||
|
if (sinp.size() != sout.size()) {
|
||||||
|
fprintf(stderr, "%s : error: input and output files have different number of tests\n", __func__);
|
||||||
|
return tests;
|
||||||
|
}
|
||||||
|
|
||||||
|
for (size_t i = 0; i < sinp.size(); ++i) {
|
||||||
|
const std::string & s = sinp[i];
|
||||||
|
const std::string & o = string_strip(sout[i]);
|
||||||
|
|
||||||
|
std::vector<llama_token> toks;
|
||||||
|
|
||||||
|
size_t pos = 0;
|
||||||
|
while (pos < o.size()) {
|
||||||
|
size_t next = o.find(' ', pos);
|
||||||
|
if (next == std::string::npos) {
|
||||||
|
next = o.size();
|
||||||
|
}
|
||||||
|
const std::string stok = o.substr(pos, next - pos);
|
||||||
|
toks.push_back(std::stoi(stok));
|
||||||
|
pos = next + 1;
|
||||||
|
}
|
||||||
|
|
||||||
|
tests[s] = toks;
|
||||||
|
}
|
||||||
|
|
||||||
|
return tests;
|
||||||
|
}
|
||||||
|
|
||||||
|
int main(int argc, char **argv) {
|
||||||
|
if (argc < 2) {
|
||||||
|
fprintf(stderr, "Usage: %s vocab-file [text-file]\n", argv[0]);
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
|
||||||
|
const std::string fname = argv[1];
|
||||||
|
|
||||||
|
const std::string fname_inp = fname + ".inp";
|
||||||
|
const std::string fname_out = fname + ".out";
|
||||||
|
|
||||||
|
std::string fname_text;
|
||||||
|
if (argc > 2) {
|
||||||
|
fname_text = argv[2];
|
||||||
|
}
|
||||||
|
|
||||||
|
fprintf(stderr, "%s : reading vocab from: '%s'\n", __func__, fname.c_str());
|
||||||
|
|
||||||
|
llama_model * model;
|
||||||
|
llama_context * ctx;
|
||||||
|
|
||||||
|
llama_backend_init();
|
||||||
|
|
||||||
|
// load the vocab
|
||||||
|
{
|
||||||
|
auto mparams = llama_model_default_params();
|
||||||
|
|
||||||
|
mparams.vocab_only = true;
|
||||||
|
|
||||||
|
model = llama_load_model_from_file(fname.c_str(), mparams);
|
||||||
|
|
||||||
|
if (model == NULL) {
|
||||||
|
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
|
||||||
|
auto cparams = llama_context_default_params();
|
||||||
|
|
||||||
|
ctx = llama_new_context_with_model(model, cparams);
|
||||||
|
|
||||||
|
if (ctx == NULL) {
|
||||||
|
fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
|
||||||
|
llama_free_model(model);
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
#ifdef _WIN32
|
||||||
|
// We need this for unicode console support
|
||||||
|
console::init(false, false);
|
||||||
|
atexit([]() { console::cleanup(); });
|
||||||
|
#endif
|
||||||
|
|
||||||
|
bool success = true;
|
||||||
|
|
||||||
|
const auto k_tests = read_tests(fname_inp, fname_out);
|
||||||
|
|
||||||
|
if (k_tests.empty()) {
|
||||||
|
fprintf(stderr, "%s : error: no tests found\n", __func__);
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
|
||||||
|
const bool add_special = false;
|
||||||
|
|
||||||
|
for (const auto & test_kv : k_tests) {
|
||||||
|
const std::vector<llama_token> res = llama_tokenize(ctx, test_kv.first, add_special);
|
||||||
|
|
||||||
|
printf("\n");
|
||||||
|
printf("src: '%s'\n", test_kv.first.c_str());
|
||||||
|
printf("res: '%s'\n", llama_detokenize_bpe(ctx, res).c_str());
|
||||||
|
printf("tok: ");
|
||||||
|
for (const auto & tok : res) {
|
||||||
|
printf("%d ", tok);
|
||||||
|
}
|
||||||
|
printf("\n");
|
||||||
|
|
||||||
|
bool correct = res.size() == test_kv.second.size();
|
||||||
|
for (int i = 0; i < (int) res.size() && correct; ++i) {
|
||||||
|
if (test_kv.second[i] != res[i]) {
|
||||||
|
correct = false;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
if (!correct) {
|
||||||
|
fprintf(stderr, "%s : failed test: '%s'\n", __func__, test_kv.first.c_str());
|
||||||
|
fprintf(stderr, "%s : detokenized to: '%s' instead of '%s'\n", __func__,
|
||||||
|
llama_detokenize_bpe(ctx, res).c_str(),
|
||||||
|
llama_detokenize_bpe(ctx, test_kv.second).c_str());
|
||||||
|
fprintf(stderr, "%s : expected tokens: ", __func__);
|
||||||
|
for (const auto & t : test_kv.second) {
|
||||||
|
fprintf(stderr, "%6d '%s', ", t, llama_token_to_piece(ctx, t).c_str());
|
||||||
|
}
|
||||||
|
fprintf(stderr, "\n");
|
||||||
|
fprintf(stderr, "%s : got tokens: ", __func__);
|
||||||
|
for (const auto & t : res) {
|
||||||
|
fprintf(stderr, "%6d '%s', ", t, llama_token_to_piece(ctx, t).c_str());
|
||||||
|
}
|
||||||
|
fprintf(stderr, "\n");
|
||||||
|
|
||||||
|
success = false;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
if (!fname_text.empty()) {
|
||||||
|
fprintf(stderr, "%s : tokenizing: '%s'\n", __func__, fname_text.c_str());
|
||||||
|
|
||||||
|
std::string text;
|
||||||
|
{
|
||||||
|
std::ifstream ifs(fname_text);
|
||||||
|
if (!ifs) {
|
||||||
|
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_text.c_str());
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
text = std::string(std::istreambuf_iterator<char>(ifs), std::istreambuf_iterator<char>());
|
||||||
|
}
|
||||||
|
|
||||||
|
fprintf(stderr, "%s : text size: %zu\n", __func__, text.size());
|
||||||
|
|
||||||
|
const std::vector<llama_token> res = llama_tokenize(ctx, text, add_special);
|
||||||
|
|
||||||
|
fprintf(stderr, "%s : tokens: %zu\n", __func__, res.size());
|
||||||
|
|
||||||
|
{
|
||||||
|
const std::string fname_out = fname_text + ".tokcpp";
|
||||||
|
|
||||||
|
std::ofstream ofs(fname_out);
|
||||||
|
if (!ofs) {
|
||||||
|
fprintf(stderr, "%s : error: could not open file '%s'\n", __func__, fname_out.c_str());
|
||||||
|
return 1;
|
||||||
|
}
|
||||||
|
|
||||||
|
for (const auto & tok : res) {
|
||||||
|
ofs << tok << " '" << string_strip(llama_detokenize_bpe(ctx, std::vector<int>{tok})) << "'" << std::endl;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
fprintf(stderr, "%s : tokens written to '%s'\n", __func__, (fname_text + ".tokcpp").c_str());
|
||||||
|
}
|
||||||
|
|
||||||
|
llama_free_model(model);
|
||||||
|
llama_free(ctx);
|
||||||
|
|
||||||
|
llama_backend_free();
|
||||||
|
|
||||||
|
printf("\n");
|
||||||
|
printf("Tests %s\n", success ? "passed" : "failed");
|
||||||
|
|
||||||
|
return success ? 0 : 3;
|
||||||
|
}
|
||||||
@ -12,7 +12,7 @@
|
|||||||
#include <thread>
|
#include <thread>
|
||||||
#include <vector>
|
#include <vector>
|
||||||
|
|
||||||
int main(int argc, char **argv) {
|
int main(int argc, char ** argv) {
|
||||||
if (argc < 2) {
|
if (argc < 2) {
|
||||||
fprintf(stderr, "Usage: %s <vocab-file>\n", argv[0]);
|
fprintf(stderr, "Usage: %s <vocab-file>\n", argv[0]);
|
||||||
return 1;
|
return 1;
|
||||||
@ -1,4 +1,4 @@
|
|||||||
#include "unicode-data.h"
|
#include "unicode-data.h"
|
||||||
|
|
||||||
#include <cstdint>
|
#include <cstdint>
|
||||||
#include <map>
|
#include <map>
|
||||||
|
|||||||
@ -12,5 +12,5 @@ extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_accent_ma
|
|||||||
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_punctuation;
|
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_punctuation;
|
||||||
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_symbol;
|
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_symbol;
|
||||||
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_control;
|
extern const std::vector<std::pair<uint32_t, uint32_t>> unicode_ranges_control;
|
||||||
extern const std::multimap<uint32_t, uint32_t> unicode_map_nfd;
|
extern const std::multimap<uint32_t, uint32_t> unicode_map_nfd;
|
||||||
extern const std::map<char32_t, char32_t> unicode_map_lowercase;
|
extern const std::map<char32_t, char32_t> unicode_map_lowercase;
|
||||||
|
|||||||
487
unicode.cpp
487
unicode.cpp
@ -5,11 +5,14 @@
|
|||||||
#include <cstddef>
|
#include <cstddef>
|
||||||
#include <cstdint>
|
#include <cstdint>
|
||||||
#include <map>
|
#include <map>
|
||||||
|
#include <regex>
|
||||||
#include <stdexcept>
|
#include <stdexcept>
|
||||||
#include <string>
|
#include <string>
|
||||||
#include <unordered_map>
|
#include <unordered_map>
|
||||||
#include <utility>
|
#include <utility>
|
||||||
#include <vector>
|
#include <vector>
|
||||||
|
#include <locale>
|
||||||
|
#include <codecvt>
|
||||||
|
|
||||||
static std::string unicode_cpts_to_utf8(const std::vector<uint32_t> & cps) {
|
static std::string unicode_cpts_to_utf8(const std::vector<uint32_t> & cps) {
|
||||||
std::string result;
|
std::string result;
|
||||||
@ -53,23 +56,22 @@ static uint32_t unicode_cpt_from_utf8(const std::string & utf8, size_t & offset)
|
|||||||
offset += 4;
|
offset += 4;
|
||||||
return result;
|
return result;
|
||||||
}
|
}
|
||||||
throw std::invalid_argument("invalid string");
|
throw std::invalid_argument("failed to convert utf8 to codepoint");
|
||||||
}
|
}
|
||||||
|
|
||||||
static std::vector<uint16_t> unicode_cpt_to_utf16(uint32_t cp) {
|
//static std::vector<uint16_t> unicode_cpt_to_utf16(uint32_t cp) {
|
||||||
std::vector<uint16_t> result;
|
// std::vector<uint16_t> result;
|
||||||
if (/* 0x0000 <= cp && */ cp <= 0xffff) {
|
// if (/* 0x0000 <= cp && */ cp <= 0xffff) {
|
||||||
result.emplace_back(cp);
|
// result.emplace_back(cp);
|
||||||
}
|
// return result;
|
||||||
else if (0x10000 <= cp && cp <= 0x10ffff) {
|
// }
|
||||||
result.emplace_back(0xd800 | ((cp - 0x10000) >> 10));
|
// if (0x10000 <= cp && cp <= 0x10ffff) {
|
||||||
result.emplace_back(0xdc00 | ((cp - 0x10000) & 0x03ff));
|
// result.emplace_back(0xd800 | ((cp - 0x10000) >> 10));
|
||||||
}
|
// result.emplace_back(0xdc00 | ((cp - 0x10000) & 0x03ff));
|
||||||
else {
|
// return result;
|
||||||
throw std::invalid_argument("invalid cpt");
|
// }
|
||||||
}
|
// throw std::invalid_argument("failed to convert codepoint to utf16");
|
||||||
return result;
|
//}
|
||||||
}
|
|
||||||
|
|
||||||
//static std::vector<uint16_t> unicode_cpts_to_utf16(const std::vector<uint32_t> & cps) {
|
//static std::vector<uint16_t> unicode_cpts_to_utf16(const std::vector<uint32_t> & cps) {
|
||||||
// std::vector<uint16_t> result;
|
// std::vector<uint16_t> result;
|
||||||
@ -80,28 +82,28 @@ static std::vector<uint16_t> unicode_cpt_to_utf16(uint32_t cp) {
|
|||||||
// return result;
|
// return result;
|
||||||
//}
|
//}
|
||||||
|
|
||||||
static uint32_t cpt_from_utf16(const std::vector<uint16_t> & utf16, size_t & offset) {
|
//static uint32_t unicode_cpt_from_utf16(const std::vector<uint16_t> & utf16, size_t & offset) {
|
||||||
assert(offset < utf16.size());
|
// assert(offset < utf16.size());
|
||||||
if (((utf16[0] >> 10) << 10) != 0xd800) {
|
// if (((utf16[0] >> 10) << 10) != 0xd800) {
|
||||||
auto result = utf16[offset + 0];
|
// auto result = utf16[offset + 0];
|
||||||
offset += 1;
|
// offset += 1;
|
||||||
return result;
|
// return result;
|
||||||
}
|
// }
|
||||||
|
//
|
||||||
if (offset + 1 >= utf16.size() || !((utf16[1] & 0xdc00) == 0xdc00)) {
|
// if (offset + 1 >= utf16.size() || !((utf16[1] & 0xdc00) == 0xdc00)) {
|
||||||
throw std::invalid_argument("invalid character");
|
// throw std::invalid_argument("invalid character");
|
||||||
}
|
// }
|
||||||
|
//
|
||||||
auto result = 0x10000 + (((utf16[0] & 0x03ff) << 10) | (utf16[1] & 0x03ff));
|
// auto result = 0x10000 + (((utf16[0] & 0x03ff) << 10) | (utf16[1] & 0x03ff));
|
||||||
offset += 2;
|
// offset += 2;
|
||||||
return result;
|
// return result;
|
||||||
}
|
//}
|
||||||
|
|
||||||
//static std::vector<uint32_t> unicode_cpts_from_utf16(const std::vector<uint16_t> & utf16) {
|
//static std::vector<uint32_t> unicode_cpts_from_utf16(const std::vector<uint16_t> & utf16) {
|
||||||
// std::vector<uint32_t> result;
|
// std::vector<uint32_t> result;
|
||||||
// size_t offset = 0;
|
// size_t offset = 0;
|
||||||
// while (offset < utf16.size()) {
|
// while (offset < utf16.size()) {
|
||||||
// result.push_back(cpt_from_utf16(utf16, offset));
|
// result.push_back(unicode_cpt_from_utf16(utf16, offset));
|
||||||
// }
|
// }
|
||||||
// return result;
|
// return result;
|
||||||
//}
|
//}
|
||||||
@ -194,34 +196,277 @@ static std::unordered_map<std::string, uint8_t> unicode_utf8_to_byte_map() {
|
|||||||
return map;
|
return map;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
static inline std::wstring unicode_wstring_from_utf8(const std::string & s) {
|
||||||
|
std::wstring_convert<std::codecvt_utf8<wchar_t>> conv;
|
||||||
|
return conv.from_bytes(s);
|
||||||
|
}
|
||||||
|
|
||||||
|
static std::vector<std::string> unicode_byte_encoding_process(const std::vector<std::string> & bpe_words) {
|
||||||
|
std::vector<std::string> bpe_encoded_words;
|
||||||
|
for (const auto & word : bpe_words) {
|
||||||
|
std::string text_utf;
|
||||||
|
auto utf_word = unicode_cpts_from_utf8(word);
|
||||||
|
for (size_t i = 0; i < utf_word.size(); ++i) {
|
||||||
|
text_utf += unicode_cpt_to_utf8(utf_word[i]);
|
||||||
|
}
|
||||||
|
|
||||||
|
std::string encoded_token;
|
||||||
|
for (char & c : text_utf) {
|
||||||
|
encoded_token += unicode_byte_to_utf8(c);
|
||||||
|
}
|
||||||
|
bpe_encoded_words.emplace_back(encoded_token);
|
||||||
|
}
|
||||||
|
return bpe_encoded_words;
|
||||||
|
}
|
||||||
|
|
||||||
|
// GPT2 system regex: 's|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+
|
||||||
|
static std::vector<size_t> unicode_regex_split_custom_gpt2(const std::string & text, const std::vector<size_t> & offsets) {
|
||||||
|
std::vector<size_t> bpe_offsets; // store the offset of each word
|
||||||
|
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
|
||||||
|
|
||||||
|
size_t start = 0;
|
||||||
|
|
||||||
|
const auto cpts = unicode_cpts_from_utf8(text);
|
||||||
|
|
||||||
|
for (auto offset : offsets) {
|
||||||
|
std::string token;
|
||||||
|
|
||||||
|
bool collecting_numeric = false;
|
||||||
|
bool collecting_letter = false;
|
||||||
|
bool collecting_special = false;
|
||||||
|
bool collecting_whitespace_lookahead = false;
|
||||||
|
bool collecting = false;
|
||||||
|
|
||||||
|
std::vector<std::string> text_utf;
|
||||||
|
text_utf.reserve(offset);
|
||||||
|
|
||||||
|
for (size_t i = start; i < start + offset; ++i) {
|
||||||
|
text_utf.emplace_back(unicode_cpt_to_utf8(cpts[i]));
|
||||||
|
}
|
||||||
|
|
||||||
|
for (int i = 0; i < (int)text_utf.size(); i++) {
|
||||||
|
const std::string & utf_char = text_utf[i];
|
||||||
|
bool split_condition = false;
|
||||||
|
int bytes_remain = text_utf.size() - i;
|
||||||
|
|
||||||
|
// forward backward lookups
|
||||||
|
const std::string & utf_char_next = (i + 1 < (int)text_utf.size()) ? text_utf[i + 1] : "";
|
||||||
|
const std::string & utf_char_next_next = (i + 2 < (int)text_utf.size()) ? text_utf[i + 2] : "";
|
||||||
|
|
||||||
|
// handling contractions
|
||||||
|
if (!split_condition && bytes_remain >= 2) {
|
||||||
|
// 's|'t|'m|'d
|
||||||
|
if (utf_char == "\'" && (utf_char_next == "s" || utf_char_next == "t" || utf_char_next == "m" || utf_char_next == "d")) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
if (split_condition) {
|
||||||
|
if (token.size()) {
|
||||||
|
bpe_offsets.emplace_back(unicode_cpts_from_utf8(token).size());
|
||||||
|
}
|
||||||
|
token = utf_char + utf_char_next;
|
||||||
|
bpe_offsets.emplace_back(unicode_cpts_from_utf8(token).size());
|
||||||
|
token = "";
|
||||||
|
i++;
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
if (!split_condition && bytes_remain >= 3) {
|
||||||
|
// 're|'ve|'ll
|
||||||
|
if (utf_char == "\'" && (
|
||||||
|
(utf_char_next == "r" && utf_char_next_next == "e") ||
|
||||||
|
(utf_char_next == "v" && utf_char_next_next == "e") ||
|
||||||
|
(utf_char_next == "l" && utf_char_next_next == "l"))
|
||||||
|
) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
if (split_condition) {
|
||||||
|
// current token + next token can be defined
|
||||||
|
if (token.size()) {
|
||||||
|
bpe_offsets.emplace_back(unicode_cpts_from_utf8(token).size());
|
||||||
|
}
|
||||||
|
token = utf_char;
|
||||||
|
token += utf_char_next;
|
||||||
|
token += utf_char_next_next;
|
||||||
|
|
||||||
|
bpe_offsets.emplace_back(unicode_cpts_from_utf8(token).size());
|
||||||
|
token = "";
|
||||||
|
i += 2;
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
if (!split_condition && !collecting) {
|
||||||
|
if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_LETTER || (token.empty() && utf_char == " " && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_LETTER)) {
|
||||||
|
collecting_letter = true;
|
||||||
|
collecting = true;
|
||||||
|
}
|
||||||
|
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_DIGIT || (token.empty() && utf_char == " " && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_DIGIT)) {
|
||||||
|
collecting_numeric = true;
|
||||||
|
collecting = true;
|
||||||
|
}
|
||||||
|
else if (
|
||||||
|
((unicode_cpt_type(utf_char) != CODEPOINT_TYPE_LETTER && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_DIGIT) && (unicode_cpt_type(utf_char) != CODEPOINT_TYPE_WHITESPACE)) ||
|
||||||
|
(token.empty() && utf_char == " " && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_LETTER && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_DIGIT && unicode_cpt_type(utf_char_next) != CODEPOINT_TYPE_WHITESPACE)
|
||||||
|
) {
|
||||||
|
collecting_special = true;
|
||||||
|
collecting = true;
|
||||||
|
}
|
||||||
|
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE && unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_WHITESPACE) {
|
||||||
|
collecting_whitespace_lookahead = true;
|
||||||
|
collecting = true;
|
||||||
|
}
|
||||||
|
else if (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
else if (!split_condition && collecting) {
|
||||||
|
if (collecting_letter && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_LETTER) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
else if (collecting_numeric && unicode_cpt_type(utf_char) != CODEPOINT_TYPE_DIGIT) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
else if (collecting_special && (unicode_cpt_type(utf_char) == CODEPOINT_TYPE_LETTER || unicode_cpt_type(utf_char) == CODEPOINT_TYPE_DIGIT || unicode_cpt_type(utf_char) == CODEPOINT_TYPE_WHITESPACE)) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
else if (collecting_whitespace_lookahead && (unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_LETTER || unicode_cpt_type(utf_char_next) == CODEPOINT_TYPE_DIGIT)) {
|
||||||
|
split_condition = true;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
if (utf_char_next == "") {
|
||||||
|
split_condition = true; // final
|
||||||
|
token += utf_char;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (split_condition) {
|
||||||
|
if (token.size()) {
|
||||||
|
bpe_offsets.emplace_back(unicode_cpts_from_utf8(token).size());
|
||||||
|
}
|
||||||
|
token = utf_char;
|
||||||
|
collecting = false;
|
||||||
|
collecting_letter = false;
|
||||||
|
collecting_numeric = false;
|
||||||
|
collecting_special = false;
|
||||||
|
collecting_whitespace_lookahead = false;
|
||||||
|
}
|
||||||
|
else {
|
||||||
|
token += utf_char;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
start += offset;
|
||||||
|
}
|
||||||
|
|
||||||
|
return bpe_offsets;
|
||||||
|
}
|
||||||
|
|
||||||
|
// use std::wregex to split the text
|
||||||
|
static std::vector<size_t> unicode_regex_split_stl(const std::wstring & wtext, const std::wstring & regex_expr, const std::vector<size_t> & offsets) {
|
||||||
|
std::wregex expr(regex_expr);
|
||||||
|
std::vector<size_t> bpe_offsets; // store the offset of each word
|
||||||
|
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
|
||||||
|
size_t start = 0;
|
||||||
|
for (auto offset : offsets) {
|
||||||
|
std::wcregex_iterator it(wtext.data() + start, wtext.data() + start + offset, expr);
|
||||||
|
std::wcregex_iterator end;
|
||||||
|
|
||||||
|
int64_t start_idx = 0;
|
||||||
|
while (it != end) {
|
||||||
|
std::wcmatch match = *it;
|
||||||
|
if (match.position() > start_idx) {
|
||||||
|
bpe_offsets.emplace_back(match.position() - start_idx);
|
||||||
|
}
|
||||||
|
bpe_offsets.emplace_back(match.length());
|
||||||
|
start_idx = match.position() + match.length();
|
||||||
|
++it;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (start_idx < (int64_t) offset) {
|
||||||
|
bpe_offsets.emplace_back(offset - start_idx);
|
||||||
|
}
|
||||||
|
start += offset;
|
||||||
|
}
|
||||||
|
|
||||||
|
return bpe_offsets;
|
||||||
|
}
|
||||||
|
|
||||||
|
// use std::regex to split the text
|
||||||
|
static std::vector<size_t> unicode_regex_split_stl(const std::string & text, const std::string & regex_expr, const std::vector<size_t> & offsets) {
|
||||||
|
std::regex expr(regex_expr);
|
||||||
|
std::vector<size_t> bpe_offsets; // store the offset of each word
|
||||||
|
bpe_offsets.reserve(offsets.size()); // Reserve memory for the approximate size
|
||||||
|
size_t start = 0;
|
||||||
|
for (auto offset : offsets) {
|
||||||
|
std::cregex_iterator it(text.data() + start, text.data() + start + offset, expr);
|
||||||
|
std::cregex_iterator end;
|
||||||
|
|
||||||
|
int64_t start_idx = 0;
|
||||||
|
while (it != end) {
|
||||||
|
std::cmatch match = *it;
|
||||||
|
if (match.position() > start_idx) {
|
||||||
|
bpe_offsets.emplace_back(match.position() - start_idx);
|
||||||
|
}
|
||||||
|
bpe_offsets.emplace_back(match.length());
|
||||||
|
start_idx = match.position() + match.length();
|
||||||
|
++it;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (start_idx < (int64_t) offset) {
|
||||||
|
bpe_offsets.emplace_back(offset - start_idx);
|
||||||
|
}
|
||||||
|
start += offset;
|
||||||
|
}
|
||||||
|
|
||||||
|
return bpe_offsets;
|
||||||
|
}
|
||||||
|
|
||||||
|
static std::vector<size_t> unicode_regex_split_custom(const std::string & text, const std::string & regex_expr, const std::vector<size_t> & offsets) {
|
||||||
|
std::vector<size_t> bpe_offsets;
|
||||||
|
|
||||||
|
(void)(text);
|
||||||
|
(void)(regex_expr);
|
||||||
|
(void)(offsets);
|
||||||
|
// TODO: this implementation is actually wrong, uncomment and run:
|
||||||
|
// make -j && ./bin/test-tokenizer-0 ../models/ggml-vocab-gpt-2.gguf
|
||||||
|
//if (regex_expr == "'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)") {
|
||||||
|
// bpe_offsets = unicode_regex_split_custom_gpt2(text, offsets);
|
||||||
|
//}
|
||||||
|
|
||||||
|
return bpe_offsets;
|
||||||
|
}
|
||||||
|
|
||||||
//
|
//
|
||||||
// interface
|
// interface
|
||||||
//
|
//
|
||||||
|
|
||||||
std::string unicode_cpt_to_utf8(uint32_t cp) {
|
std::string unicode_cpt_to_utf8(uint32_t cp) {
|
||||||
std::string result;
|
std::string result;
|
||||||
|
|
||||||
if (/* 0x00 <= cp && */ cp <= 0x7f) {
|
if (/* 0x00 <= cp && */ cp <= 0x7f) {
|
||||||
result.push_back(cp);
|
result.push_back(cp);
|
||||||
|
return result;
|
||||||
}
|
}
|
||||||
else if (0x80 <= cp && cp <= 0x7ff) {
|
if (0x80 <= cp && cp <= 0x7ff) {
|
||||||
result.push_back(0xc0 | ((cp >> 6) & 0x1f));
|
result.push_back(0xc0 | ((cp >> 6) & 0x1f));
|
||||||
result.push_back(0x80 | (cp & 0x3f));
|
result.push_back(0x80 | (cp & 0x3f));
|
||||||
|
return result;
|
||||||
}
|
}
|
||||||
else if (0x800 <= cp && cp <= 0xffff) {
|
if (0x800 <= cp && cp <= 0xffff) {
|
||||||
result.push_back(0xe0 | ((cp >> 12) & 0x0f));
|
result.push_back(0xe0 | ((cp >> 12) & 0x0f));
|
||||||
result.push_back(0x80 | ((cp >> 6) & 0x3f));
|
result.push_back(0x80 | ((cp >> 6) & 0x3f));
|
||||||
result.push_back(0x80 | (cp & 0x3f));
|
result.push_back(0x80 | (cp & 0x3f));
|
||||||
|
return result;
|
||||||
}
|
}
|
||||||
else if (0x10000 <= cp && cp <= 0x10ffff) {
|
if (0x10000 <= cp && cp <= 0x10ffff) {
|
||||||
result.push_back(0xf0 | ((cp >> 18) & 0x07));
|
result.push_back(0xf0 | ((cp >> 18) & 0x07));
|
||||||
result.push_back(0x80 | ((cp >> 12) & 0x3f));
|
result.push_back(0x80 | ((cp >> 12) & 0x3f));
|
||||||
result.push_back(0x80 | ((cp >> 6) & 0x3f));
|
result.push_back(0x80 | ((cp >> 6) & 0x3f));
|
||||||
result.push_back(0x80 | (cp & 0x3f));
|
result.push_back(0x80 | (cp & 0x3f));
|
||||||
|
return result;
|
||||||
}
|
}
|
||||||
else {
|
|
||||||
throw std::invalid_argument("invalid codepoint");
|
throw std::invalid_argument("invalid codepoint");
|
||||||
}
|
|
||||||
return result;
|
|
||||||
}
|
}
|
||||||
|
|
||||||
std::vector<uint32_t> unicode_cpts_normalize_nfd(const std::vector<uint32_t> & cpts) {
|
std::vector<uint32_t> unicode_cpts_normalize_nfd(const std::vector<uint32_t> & cpts) {
|
||||||
@ -275,3 +520,167 @@ char32_t unicode_tolower(char32_t cp) {
|
|||||||
auto it = unicode_map_lowercase.find(cp);
|
auto it = unicode_map_lowercase.find(cp);
|
||||||
return it == unicode_map_lowercase.end() ? cp : it->second;
|
return it == unicode_map_lowercase.end() ? cp : it->second;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
std::vector<std::string> unicode_regex_split(const std::string & text, const std::vector<std::string> & regex_exprs) {
|
||||||
|
// unicode categories
|
||||||
|
static const std::map<std::string, int> k_ucat_enum = {
|
||||||
|
{ "\\p{N}", CODEPOINT_TYPE_DIGIT },
|
||||||
|
{ "\\p{L}", CODEPOINT_TYPE_LETTER },
|
||||||
|
{ "\\p{P}", CODEPOINT_TYPE_PUNCTUATION },
|
||||||
|
};
|
||||||
|
|
||||||
|
static const std::map<int, int> k_ucat_cpt = {
|
||||||
|
{ CODEPOINT_TYPE_DIGIT, 0xD1 },
|
||||||
|
{ CODEPOINT_TYPE_LETTER, 0xD2 },
|
||||||
|
{ CODEPOINT_TYPE_PUNCTUATION, 0xD3 },
|
||||||
|
};
|
||||||
|
|
||||||
|
static const std::map<int, std::string> k_ucat_map = {
|
||||||
|
{ CODEPOINT_TYPE_DIGIT, "\x30-\x39" }, // 0-9
|
||||||
|
{ CODEPOINT_TYPE_LETTER, "\x41-\x5A\x61-\x7A" }, // A-Za-z
|
||||||
|
{ CODEPOINT_TYPE_PUNCTUATION, "\x21-\x23\x25-\x2A\x2C-\x2F\x3A-\x3B\x3F-\x40\\\x5B-\\\x5D\x5F\\\x7B\\\x7D" }, // !-#%-*,-/:-;?-@\[-\]_\{\}
|
||||||
|
};
|
||||||
|
|
||||||
|
// compute collapsed codepoints only if needed by at least one regex
|
||||||
|
bool need_collapse = false;
|
||||||
|
for (auto & regex_expr : regex_exprs) {
|
||||||
|
// search for unicode categories
|
||||||
|
for (const auto & ucat : k_ucat_enum) {
|
||||||
|
if (std::string::npos != regex_expr.find(ucat.first)) {
|
||||||
|
need_collapse = true;
|
||||||
|
break;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
const auto cpts = unicode_cpts_from_utf8(text);
|
||||||
|
|
||||||
|
// generate a "collapsed" representation of the text, where all codepoints are replaced by a single byte
|
||||||
|
// ref: https://github.com/ggerganov/llama.cpp/pull/6920#issuecomment-2081479935
|
||||||
|
std::string text_collapsed;
|
||||||
|
if (need_collapse) {
|
||||||
|
// collapse all unicode categories
|
||||||
|
text_collapsed.resize(cpts.size());
|
||||||
|
|
||||||
|
for (size_t i = 0; i < cpts.size(); ++i) {
|
||||||
|
// keep single-byte codepoints as is
|
||||||
|
if (cpts[i] < 128) {
|
||||||
|
text_collapsed[i] = cpts[i];
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
|
||||||
|
const int cpt_type = unicode_cpt_type(cpts[i]);
|
||||||
|
|
||||||
|
if (k_ucat_cpt.find(cpt_type) != k_ucat_cpt.end()) {
|
||||||
|
text_collapsed[i] = k_ucat_cpt.at(cpt_type);
|
||||||
|
} else {
|
||||||
|
text_collapsed[i] = (char) 0xD0; // fallback
|
||||||
|
}
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
std::vector<size_t> bpe_offsets = { cpts.size() };
|
||||||
|
|
||||||
|
for (auto & regex_expr : regex_exprs) {
|
||||||
|
// first, see if we have an efficient custom regex implementation
|
||||||
|
auto tmp = unicode_regex_split_custom(text, regex_expr, bpe_offsets);
|
||||||
|
|
||||||
|
if (!tmp.empty()) {
|
||||||
|
bpe_offsets = std::move(tmp);
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
|
||||||
|
// fallback to general-purpose std::regex / std::wregex
|
||||||
|
try {
|
||||||
|
// if a unicode category is used in the regex, we use the collapsed text and replace the unicode category
|
||||||
|
// with the corresponding collapsed representation
|
||||||
|
bool use_collapsed = false;
|
||||||
|
for (auto & ucat : k_ucat_enum) {
|
||||||
|
if (std::string::npos != regex_expr.find(ucat.first)) {
|
||||||
|
use_collapsed = true;
|
||||||
|
break;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
if (use_collapsed) {
|
||||||
|
// sanity-check that the original regex does not contain any non-ASCII characters
|
||||||
|
const auto cpts_regex = unicode_cpts_from_utf8(regex_expr);
|
||||||
|
for (size_t i = 0; i < cpts_regex.size(); ++i) {
|
||||||
|
if (cpts_regex[i] >= 128) {
|
||||||
|
throw std::runtime_error("Regex includes both unicode categories and non-ASCII characters - not supported");
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
// generate a collapsed representation of the regex
|
||||||
|
std::string regex_expr_collapsed;
|
||||||
|
|
||||||
|
// track if we are inside [], because nested [] are not allowed
|
||||||
|
bool inside = false;
|
||||||
|
for (size_t i = 0; i < regex_expr.size(); ++i) {
|
||||||
|
if (regex_expr[i] == '[' && (i == 0 || regex_expr[i - 1] != '\\')) {
|
||||||
|
regex_expr_collapsed += '[';
|
||||||
|
inside = true;
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (inside && regex_expr[i] == ']' && regex_expr[i - 1] != '\\') {
|
||||||
|
regex_expr_collapsed += ']';
|
||||||
|
inside = false;
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
|
||||||
|
if (regex_expr[i + 0] == '\\' && i + 4 < regex_expr.size() &&
|
||||||
|
regex_expr[i + 1] == 'p' &&
|
||||||
|
regex_expr[i + 2] == '{' &&
|
||||||
|
regex_expr[i + 4] == '}') {
|
||||||
|
const std::string pat = regex_expr.substr(i, 5);
|
||||||
|
if (k_ucat_enum.find(pat) != k_ucat_enum.end()) {
|
||||||
|
if (!inside) {
|
||||||
|
regex_expr_collapsed += '[';
|
||||||
|
}
|
||||||
|
regex_expr_collapsed += k_ucat_cpt.at(k_ucat_enum.at(pat));
|
||||||
|
regex_expr_collapsed += k_ucat_map.at(k_ucat_enum.at(pat));
|
||||||
|
if (!inside) {
|
||||||
|
regex_expr_collapsed += ']';
|
||||||
|
}
|
||||||
|
i += 4;
|
||||||
|
continue;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
regex_expr_collapsed += regex_expr[i];
|
||||||
|
}
|
||||||
|
|
||||||
|
//printf("text_collapsed: %s\n", text_collapsed.c_str());
|
||||||
|
//printf("regex_expr_collapsed: %s\n", regex_expr_collapsed.c_str());
|
||||||
|
bpe_offsets = unicode_regex_split_stl(text_collapsed, regex_expr_collapsed, bpe_offsets);
|
||||||
|
} else {
|
||||||
|
// no unicode category used, we can use std::wregex directly
|
||||||
|
const std::wstring wtext = unicode_wstring_from_utf8(text);
|
||||||
|
const std::wstring wregex_expr = unicode_wstring_from_utf8(regex_expr);
|
||||||
|
|
||||||
|
//printf("text: %s\n", text.c_str());
|
||||||
|
//printf("regex_expr: %s\n", regex_expr.c_str());
|
||||||
|
bpe_offsets = unicode_regex_split_stl(wtext, wregex_expr, bpe_offsets);
|
||||||
|
}
|
||||||
|
} catch (std::regex_error & e) {
|
||||||
|
fprintf(stderr, "Failed to process regex: '%s'\n", regex_expr.c_str());
|
||||||
|
fprintf(stderr, "Regex error: %s\n", e.what());
|
||||||
|
throw std::runtime_error("Failed to process regex");
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
std::vector<std::string> bpe_words;
|
||||||
|
bpe_words.reserve(bpe_offsets.size()); // reserve memory for the approximate size
|
||||||
|
|
||||||
|
size_t start = 0;
|
||||||
|
for (size_t & offset : bpe_offsets) {
|
||||||
|
bpe_words.emplace_back();
|
||||||
|
for (size_t i = start; i < start + offset; ++i) {
|
||||||
|
bpe_words.back() += unicode_cpt_to_utf8(cpts[i]);
|
||||||
|
}
|
||||||
|
start += offset;
|
||||||
|
}
|
||||||
|
|
||||||
|
return unicode_byte_encoding_process(bpe_words);
|
||||||
|
}
|
||||||
|
|||||||
@ -24,5 +24,6 @@ int unicode_cpt_type(const std::string & utf8);
|
|||||||
std::string unicode_byte_to_utf8(uint8_t byte);
|
std::string unicode_byte_to_utf8(uint8_t byte);
|
||||||
uint8_t unicode_utf8_to_byte(const std::string & utf8);
|
uint8_t unicode_utf8_to_byte(const std::string & utf8);
|
||||||
|
|
||||||
// simple tolower that only implements one-to-one mapping, not one-to-many
|
|
||||||
char32_t unicode_tolower(char32_t cp);
|
char32_t unicode_tolower(char32_t cp);
|
||||||
|
|
||||||
|
std::vector<std::string> unicode_regex_split(const std::string & text, const std::vector<std::string> & regex_exprs);
|
||||||
|
|||||||
Loading…
Reference in New Issue
Block a user