llama.cpp/convert-gptneox-hf-to-gguf.py

269 lines
8.7 KiB
Python
Raw Normal View History

# HF gptneox--> gguf conversion
2023-07-30 13:05:37 +00:00
import gguf
import os
2023-07-30 13:05:37 +00:00
import sys
import struct
import json
import numpy as np
2023-08-16 16:55:49 +00:00
import torch
2023-07-30 13:05:37 +00:00
from typing import Any, List
from pathlib import Path
from transformers import AutoTokenizer
# ref: https://github.com/openai/gpt-2/blob/master/src/encoder.py
def bytes_to_unicode():
"""
Returns list of utf-8 byte and a corresponding list of unicode strings.
The reversible bpe codes work on unicode strings.
This means you need a large # of unicode characters in your vocab if you want to avoid UNKs.
When you're at something like a 10B token dataset you end up needing around 5K for decent coverage.
This is a significant percentage of your normal, say, 32K bpe vocab.
To avoid that, we want lookup tables between utf-8 bytes and unicode strings.
And avoids mapping to whitespace/control characters the bpe code barfs on.
"""
bs = list(range(ord("!"), ord("~")+1))+list(range(ord("¡"), ord("¬")+1))+list(range(ord("®"), ord("ÿ")+1))
cs = bs[:]
n = 0
for b in range(2**8):
if b not in bs:
bs.append(b)
cs.append(2**8+n)
n += 1
cs = [chr(n) for n in cs]
return dict(zip(bs, cs))
2023-07-30 13:05:37 +00:00
def count_model_parts(dir_model: str) -> int:
num_parts = 0
for filename in os.listdir(dir_model):
if filename.startswith("pytorch_model-"):
num_parts += 1
if num_parts > 0:
print("gguf: found " + str(num_parts) + " model parts")
return num_parts
2023-07-30 13:05:37 +00:00
if len(sys.argv) < 3:
print("Usage: convert-h5-to-ggml.py dir-model ftype\n")
print(" ftype == 0 -> float32")
print(" ftype == 1 -> float16")
sys.exit(1)
# output in the same directory as the model
dir_model = sys.argv[1]
last_dir = os.path.basename(os.path.normpath(dir_model))
2023-07-30 13:05:37 +00:00
# possible tensor data types
# ftype == 0 -> float32
# ftype == 1 -> float16
2023-07-30 13:05:37 +00:00
# map from ftype to string
ftype_str = ["f32", "f16"]
ftype = 1
if len(sys.argv) > 2:
ftype = int(sys.argv[2])
if ftype < 0 or ftype > 1:
print("Invalid ftype: " + str(ftype))
2023-07-30 13:05:37 +00:00
sys.exit(1)
fname_out = sys.argv[1] + "/ggml-model-" + ftype_str[ftype] + ".gguf"
2023-07-30 13:05:37 +00:00
print("gguf: loading model "+last_dir)
2023-07-31 01:00:20 +00:00
with open(dir_model + "/config.json", "r", encoding="utf-8") as f:
hparams = json.load(f)
if hparams["architectures"][0] != "GPTNeoXForCausalLM":
print("Model architecture not supported: " + hparams["architectures"][0])
2023-07-31 01:00:20 +00:00
sys.exit()
2023-07-30 13:05:37 +00:00
# get number of model parts
num_parts = count_model_parts(dir_model)
2023-07-30 13:05:37 +00:00
llm_arch = "gptneox"
gguf_writer = gguf.GGUFWriter(fname_out, arch=llm_arch)
2023-07-30 13:05:37 +00:00
print("gguf: get model metadata")
2023-07-30 13:05:37 +00:00
block_count = hparams["num_hidden_layers"]
2023-07-30 13:05:37 +00:00
gguf_writer.add_architecture()
gguf_writer.add_name(last_dir)
gguf_writer.add_context_length(hparams["max_position_embeddings"])
gguf_writer.add_embedding_length(hparams["hidden_size"])
gguf_writer.add_block_count(block_count)
gguf_writer.add_feed_forward_length(hparams["intermediate_size"])
gguf_writer.add_rope_dimension_count(int(hparams["rotary_pct"]*(hparams["hidden_size"]//hparams["num_attention_heads"])))
gguf_writer.add_head_count(hparams["num_attention_heads"])
gguf_writer.add_parallel_residual(hparams["use_parallel_residual"] if "use_parallel_residual" in hparams else True)
gguf_writer.add_layer_norm_eps(hparams["layer_norm_eps"])
2023-07-30 13:05:37 +00:00
# TOKENIZATION
print("gguf: get tokenizer metadata")
2023-07-30 13:05:37 +00:00
tokens: List[str] = []
merges: List[str] = []
2023-07-30 13:05:37 +00:00
if Path(dir_model + "/tokenizer.json").is_file():
# gpt2 tokenizer
gguf_writer.add_tokenizer_model("gpt2")
print("gguf: get gpt2 tokenizer merges")
2023-07-30 13:05:37 +00:00
with open(dir_model + "/tokenizer.json", "r", encoding="utf-8") as f:
tokenizer_json = json.load(f)
merges = tokenizer_json["model"]["merges"]
2023-07-30 13:05:37 +00:00
gguf_writer.add_token_merges(merges)
2023-07-30 13:05:37 +00:00
print("gguf: get gpt2 tokenizer vocab")
vocab_size = len(tokenizer_json["model"]["vocab"])
# ref: https://github.com/cmp-nct/ggllm.cpp/blob/master/falcon_convert.py
tokenizer = AutoTokenizer.from_pretrained(dir_model)
reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()}
byte_encoder = bytes_to_unicode()
byte_decoder = {v: k for k, v in byte_encoder.items()}
for i in range(vocab_size):
if i in reverse_vocab:
try:
text = bytearray([byte_decoder[c] for c in reverse_vocab[i]])
except KeyError:
text = bytearray()
for c in reverse_vocab[i]:
if ord(c) < 256: # single byte character
text.append(byte_decoder[ord(c)])
else: # multibyte special token character
text.extend(c.encode('utf-8'))
else:
print(f"Key {i} not in tokenizer vocabulary. Padding with an arbitrary token.")
pad_token = f"[PAD{i}]".encode("utf8")
text = bytearray(pad_token)
tokens.append(text)
2023-07-30 13:05:37 +00:00
gguf_writer.add_token_list(tokens)
if "added_tokens" in tokenizer_json and Path(dir_model + "/tokenizer_config.json").is_file():
print("gguf: get special token ids")
with open(dir_model + "/tokenizer_config.json", "r", encoding="utf-8") as f:
tokenizer_config = json.load(f)
# find special token ids
if "bos_token" in tokenizer_config:
for key in tokenizer_json["added_tokens"]:
if key["content"] == tokenizer_config["bos_token"]:
gguf_writer.add_bos_token_id(key["id"])
if "eos_token" in tokenizer_config:
for key in tokenizer_json["added_tokens"]:
if key["content"] == tokenizer_config["eos_token"]:
gguf_writer.add_eos_token_id(key["id"])
if "unk_token" in tokenizer_config:
for key in tokenizer_json["added_tokens"]:
if key["content"] == tokenizer_config["unk_token"]:
gguf_writer.add_unk_token_id(key["id"])
if "sep_token" in tokenizer_config:
for key in tokenizer_json["added_tokens"]:
if key["content"] == tokenizer_config["sep_token"]:
gguf_writer.add_sep_token_id(key["id"])
if "pad_token" in tokenizer_config:
for key in tokenizer_json["added_tokens"]:
if key["content"] == tokenizer_config["pad_token"]:
gguf_writer.add_pad_token_id(key["id"])
2023-07-30 13:05:37 +00:00
# TENSORS
2023-08-16 16:55:49 +00:00
tensor_map = gguf.get_tensor_name_map(block_count)
2023-07-30 13:05:37 +00:00
# tensor info
print("gguf: get tensor metadata")
2023-07-30 13:05:37 +00:00
if num_parts == 0:
part_names = ("pytorch_model.bin",)
else:
part_names = (
f"pytorch_model-{n:05}-of-{num_parts:05}.bin" for n in range(1, num_parts + 1)
)
2023-07-30 13:05:37 +00:00
for part_name in part_names:
print("gguf: loading model part '" + part_name + "'")
model_part = torch.load(f"{dir_model}/{part_name}", map_location="cpu")
for name in model_part.keys():
data = model_part[name]
2023-07-30 13:05:37 +00:00
# we don't need these
if name.endswith(".attention.masked_bias") or name.endswith(".attention.bias") or name.endswith(".attention.rotary_emb.inv_freq"):
continue
# convert any unsupported data types to float32
if data.dtype != torch.float16 and data.dtype != torch.float32:
data = data.to(torch.float32)
data = data.squeeze().numpy()
# map tensor names
if name.endswith(".weight") and name[:-7] in tensor_map:
name = tensor_map[name[:-7]] + ".weight"
elif name.endswith(".bias") and name[:-5] in tensor_map:
name = tensor_map[name[:-5]] + ".bias"
else:
print("Can not map tensor '" + name + "'")
sys.exit()
n_dims = len(data.shape)
2023-08-16 16:55:49 +00:00
data_dtype = data.dtype
old_dtype = data_dtype
# if f32 desired, convert any float16 to float32
if ftype == 0 and data.dtype == np.float16:
data_dtype = np.float32
# TODO: Why cant we use these float16 as-is? There should be not reason to store float16 as float32
if ftype == 1 and data.dtype == np.float16 and n_dims == 1:
data_dtype = np.float32
# if f16 desired, convert any float32 2-dim weight tensors to float16
if ftype == 1 and data.dtype == np.float32 and name.endswith(".weight") and n_dims == 2:
data_dtype = np.float16
print(name + ", n_dims = " + str(n_dims) + ", " + str(old_dtype) + " --> " + str(data_dtype))
data = data.astype(data_dtype)
gguf_writer.add_tensor(name, data)
2023-07-30 13:05:37 +00:00
print("gguf: write header")
gguf_writer.write_header_to_file()
print("gguf: write metadata")
gguf_writer.write_kv_data_to_file()
print("gguf: write tensors")
gguf_writer.write_tensors_to_file()
2023-07-30 13:05:37 +00:00
gguf_writer.close()
print("gguf: model successfully exported to '" + fname_out + "'")
2023-07-30 13:05:37 +00:00
print("")