Commit Graph

51 Commits

Author SHA1 Message Date
tjohnman
305ba6f0e6
Don't force immediate interactive without -i (#354)
* Don't force immediate interactive without -i

Sometimes we might want to use a reverse prompt but we want to let the
model generate tokens right after the initial prompt. So we don't force
user input mode if the -i flag wasn't specified and instead let it run
until we encounter the reverse prompt.

This gives use some more flexibility, since it doesn't force the user to
enter a newline if they want to let the model generate text right after
the initial prompt and only be asked for input if the reverse prompt is
encountered.

The `--interactive-first` flag is reintroduced to force the old
behavior. `-r` behaves like `-i` plus introduces a reverse prompt (it
can be specified more than once).

* Update help output.

---------

Co-authored-by: Johnman <tjohnman@github>
2023-03-22 19:16:35 +02:00
Erik Scholz
56e659a0b2
fix perplexity after c-api refactor (#390)
* preallocate a buffer of fitting size for tokenization (utils.cpp)

* don't create a new std::string (especially here, where it's usually large)
2023-03-22 18:09:38 +02:00
Georgi Gerganov
ae44e23ee3
When seed <= 0 - use the clock to generate one 2023-03-22 07:47:15 +02:00
Georgi Gerganov
928480ef5b
Init llama_context_params properly from CLI (#370) 2023-03-22 07:45:14 +02:00
Georgi Gerganov
f5a77a629b
Introduce C-style API (#370)
* Major refactoring - introduce C-style API

* Clean up

* Add <cassert>

* Add <iterator>

* Add <algorithm> ....

* Fix timing reporting and accumulation

* Measure eval time only for single-token calls

* Change llama_tokenize return meaning
2023-03-22 07:32:36 +02:00
Fabio R. Sluzala
353ec251a4
We could use std::unordered_map over std::map (#305)
* Improve performance by changing std::map to std::unordered_map and std::map<id, token> id_to_token; to std::vector<token> id_to_token;

* fix last commit on gpt_vocab_init add vocab.id_to_token.resize(vocab.token_to_id.size());

* Removed include <map>

* Nest struct token score inside gpt_vocab

* renamed token to tok
2023-03-21 19:21:50 +02:00
Matvey Soloviev
89d5d90f3b
Fix color codes emitting mid-UTF8 code. (#312) 2023-03-21 19:11:01 +02:00
comex
16ffc013c6
Importer for GPTQ quantized LLaMA models (#301)
* [WIP, broken] Importer for GPTQ quantized LLaMA models

Based on: https://github.com/qwopqwop200/GPTQ-for-LLaMa

Current status: Something is busted.  The output starts out decent, but
quickly degrades into gibberish.  This doesn't happen with either the
original GPTQ-for-LLaMa using the same weights, or llama.cpp when using
weights quantized by its own quantizer.  Is there a bug in the
conversion script that somehow only comes into play with a large context
size?

I did notice one potential issue.  It's clearly not the main cause of
the gibberish, since it doesn't happen when using q4_1 weights quantized
by llama.cpp itself, but it seems concerning.  When doing a matrix
multiplication of f16 * f32 => f32 or q4_1 * f32 => f32, at least when
the multiplication is not done with BLAS, the intermediate results are
stored in the smaller format rather than f32.  This seems like an
unnecessary waste of precision, especially in the q4_1 case.

I was originally hoping to validate the results by matching the Python
implementation's output exactly, but precision and non-associativity
issues make this very difficult, including when performing matrix
multiplications and, especially, computing norms.

Anyway, design details:

The models being imported store per-layer weights in essentially q4_1
format, although the addend and scale are shared across an entire row
rather than every group of 32 weights.  This script duplicates the
addend and scale to match ggml's expectations, at the cost of wasting
some memory.

However, there are two differences which I accommodated changing the
output format (and adding corresponding support to main.cpp) rather than
having the script match the existing one:

- The tok_embeddings and output weights (i.e. the weights that aren't
  per-layer) are f16 instead of q4_1.  They could be converted to q4_1,
  and the impact of the loss of precision would probably be low, but
  this would rule out exactly matching the Python implementation's
  output for validation.

- There is no sharding, since the input doesn't have it, and for a
  CPU-only implementation it seems more useful to avoid having to deal
  with multiple files.

The new format is differentiated from existing q4_1 format by changing
the 'f16' header flag to a new value, 4.  That said, I think a cleaner
approach would be to change main.cpp to support loading each tensor with
an arbitrary sharding configuration and type rather than hardcoding
specific combinations of types.  So far I've wasted too much time
debugging to try implementing this...

* Add missing permutation.  Now it works.

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-21 18:42:25 +02:00
Gary Linscott
486ae645fd
Compute perplexity over prompt (#270)
* Compute perplexity over prompt

* More accurate perplexity calculation - over all logits in the context window (so 512x more tokens!)

* Output all perplexitiies

* Add timing/ETA
2023-03-21 18:27:42 +02:00
anzz1
c86ba036e6
Enable ANSI colors on Windows 10+ (#311)
* Enable ANSI colors on Windows 10+

On older versions function will silently fail without any ill effects

* Do not call SetConsoleMode if the mode is already set

* Update main.cpp

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-21 18:14:46 +02:00
tjohnman
d5f56a5e5a
Check for reverse prompt by characters instead of tokens (#292) (#330)
* Check for reverse prompt by characters instead of tokens (#292)

* Update main.cpp

Wording.

* Cleanup.

* Remove unnecessary use of std::stringstream.

---------

Co-authored-by: Johnman <tjohnman@github>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-21 18:04:43 +02:00
Georgi Gerganov
3bfa3b43b7
Fix convert script, warnings alpaca instructions, default params 2023-03-21 17:59:16 +02:00
anzz1
975d2cebf9
cmdline option for custom amount of model parts (--n_parts N) (#348)
* cmdline option for custom amount of model parts (--n_parts N)

* Update main.cpp

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-21 17:42:43 +02:00
Georgi Gerganov
eb34620aec
Add tokenizer test + revert to C++11 (#355)
* Add test-tokenizer-0 to do a few tokenizations - feel free to expand
* Added option to convert-pth-to-ggml.py script to dump just the vocabulary
* Added ./models/ggml-vocab.bin containing just LLaMA vocab data (used for tests)
* Added utility to load vocabulary file from previous point (temporary implementation)
* Avoid using std::string_view and drop back to C++11 (hope I didn't break something)
* Rename gpt_vocab -> llama_vocab
* All CMake binaries go into ./bin/ now
2023-03-21 17:29:41 +02:00
Mack Straight
a791a68b61
move file magic/version to header, print expected version (#319) 2023-03-20 19:26:01 +00:00
Mack Straight
074bea2eb1
sentencepiece bpe compatible tokenizer (#252)
* potential out of bounds read

* fix quantize

* style

* Update convert-pth-to-ggml.py

* mild cleanup

* don't need the space-prefixing here rn since main.cpp already does it

* new file magic + version header field

* readme notice

* missing newlines

Co-authored-by: slaren <2141330+slaren@users.noreply.github.com>
2023-03-20 03:17:23 -07:00
cocktailpeanut
da5303c1ea
bugfix: default should not be interactive (#304) 2023-03-19 23:44:20 +02:00
Rickey Bowers Jr
5c19c70ba6
fix coloring of last n_batch of prompt, and refactor line input (#221)
* fix coloring of last `n_batch` of prompt, and refactor line input
* forgot the newline that needs to be sent to the model
* (per #283) try to force flush of color reset in SIGINT handler
2023-03-19 19:44:30 +00:00
tjohnman
24568371ae
Support for multiple reverse prompts. (#299)
Co-authored-by: Johnman <>
Co-authored-by: Johnman <tjohnman@github>
2023-03-19 21:33:06 +02:00
tjohnman
ad5fd5b60c
Make prompt randomization optional. (#300)
Co-authored-by: Johnman <>
2023-03-19 20:36:19 +02:00
tjohnman
368d0c8a9e
Respect the maximum number of tokens in interactive. (#298)
Co-authored-by: Johnman <johnman@github>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-19 20:31:17 +02:00
slaren
50fae10d03
Add --ignore-eos parameter (#181)
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-19 20:22:48 +02:00
Qingyou Meng
084e2f0ec0
interactive mode: print '\n' in sigint_handler, this flush stdout thus ensure color reset. (#283) 2023-03-19 20:10:00 +02:00
Erik Scholz
0b366e7357
Command line switch to use F16 for memory_k and memory_v (refactor of #154) (#294)
* Use F16 for memory_k and memory_v

* add command line switch to use f16 instead of f32 for memory k+v

---------

Co-authored-by: Ty Everett <ty@tyweb.us>
2023-03-19 19:57:00 +02:00
Georgi Gerganov
c494ed5b94
Fix off-by-one bug (#115) 2023-03-19 19:46:32 +02:00
Georgi Gerganov
70f01cb863
Drop trailing new line from file prompts (#80) 2023-03-19 19:05:04 +02:00
Georgi Gerganov
9e1707218a
Add "--instruct" argument for usage with Alpaca (#240)
Also start adding prompts in "./prompts"
2023-03-19 18:37:02 +02:00
Ronsor
d7def1a752
Warn user if a context size greater than 2048 tokens is specified (#274)
LLaMA doesn't support more than 2048 token context sizes, and going above that produces terrible results.
2023-03-18 20:10:47 -04:00
Alex Nguyen
d3f202d57b
Remove unused code since n_vocab is model.hparams.n_vocab (#262) 2023-03-18 13:51:49 +00:00
Justin Suess
e03e359730
fixed warning with std::ignore about unused function result (#151)
fixed warning with std::ignore about unused function result
2023-03-18 11:44:09 +00:00
thement
c9f670a177
Implement non-greedy tokenizer that tries to maximize token lengths (#242)
* Implement non-greedy tokenizer that tries to maximize token lengths

* Insert single space in front of the prompt

- this is to match original llama tokenizer behavior

---------

Co-authored-by: Jakub Horak <jakub.horak@ibawizard.net>
2023-03-17 21:05:58 +01:00
hoangmit
6eac39ba95
Add RMS norm and use it (#187)
* add ggml_rms_norm

* update op num
2023-03-16 00:41:38 +02:00
Rickey Bowers Jr
2d15d6c9a9
add SIGINT support for _WIN32 environments (#120)
* add SIGINT support for _WIN32 environments

* perhaps more consistent
2023-03-15 21:56:24 +02:00
Justin Suess
2d64715ad4
added ctx_size parameter (#148)
* added ctx_size parameter

* added it in more places

* Apply suggestions from code review

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-15 21:42:40 +02:00
Justin Suess
16b2c61a22
fixed color reset on exit (#149)
* fixed color reset on exit

* added sigint handler for ansi_color_reset

* Update main.cpp

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-15 21:39:38 +02:00
Georgi Gerganov
4497ad819c
Print system information 2023-03-13 19:15:08 +02:00
Pavol Rusnak
671d5cac15
Use fprintf for diagnostic output (#48)
keep printf only for printing model output

one can now use ./main ... 2>dev/null to suppress any diagnostic output
2023-03-13 18:39:56 +02:00
uint256_t
63fd76fbb0
Reduce model loading time (#43)
* Use buffering

* Use vector

* Minor

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-13 18:33:43 +02:00
Val Kharitonov
2a20f48efa
Fix UTF-8 handling (including colors) (#79) 2023-03-13 18:24:18 +02:00
Matvey Soloviev
a169bb889c Gate signal support on being on a unixoid system. (#74) 2023-03-13 04:08:01 +01:00
Matvey Soloviev
460c482540 Fix token count accounting 2023-03-13 01:04:41 +01:00
Matvey Soloviev
404fac0d62
Fix color getting reset before prompt output done (#65)
(cherry picked from commit 7eb2987619feee04c40eff69b604017d09919cb6)
2023-03-13 00:07:34 +02:00
Matvey Soloviev
96ea727f47
Add interactive mode (#61)
* Initial work on interactive mode.

* Improve interactive mode. Make rev. prompt optional.

* Update README to explain interactive mode.

* Fix OS X build
2023-03-12 23:13:28 +02:00
beiller
02f0c6fe7f
Add back top_k (#56)
* Add back top_k

* Update utils.cpp

* Update utils.h

---------

Co-authored-by: Bill Hamilton <bill.hamilton@shopify.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-12 22:23:15 +02:00
Sebastián A
eb062bb012
Windows fixes (#31)
* Apply fixes suggested to build on windows

Issue: https://github.com/ggerganov/llama.cpp/issues/22

* Remove unsupported VLAs

* MSVC: Remove features that are only available on MSVC C++20.

* Fix zero initialization of the other fields.

* Change the use of vector for stack allocations.
2023-03-12 22:15:00 +02:00
beiller
129c7d1ea8
Add repetition penalty (#20)
* Adding repeat penalization

* Update utils.h

* Update utils.cpp

* Numeric fix

Should probably still scale by temp even if penalized

* Update comments, more proper application

I see that numbers can go negative so a fix from a referenced commit

* Minor formatting

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2023-03-12 11:27:42 +02:00
Georgi Gerganov
7d9ed7b25f
Bump memory buffer 2023-03-11 12:45:01 +02:00
Georgi Gerganov
007a8f6f45
Support all LLaMA models + change Q4_0 quantization storage 2023-03-11 11:28:30 +02:00
Georgi Gerganov
70bc0b8b15
Fix a bug in the rope calculation 2023-03-10 23:46:57 +02:00
Georgi Gerganov
319cdb3e1f
Final touches 2023-03-10 21:50:46 +02:00