423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-18 12:50:38 +08:00
Code Issues Packages Projects Releases Wiki Activity
2,645 Commits 1 Branch 34 Tags
c31c63b41109e616997757ec2da6e0ab89ed3b6e
Commit Graph

732 Commits

Author SHA1 Message Date
hoshi-hiyouga
c31c63b411 [misc] fix grad ckpt (#6931) 2025-02-13 23:27:51 +08:00
hoshi-hiyouga
797043d29c [model] add liger kernel to qwen2_5 vl (#6930) 
* add liger kernel to qwen2_5 vl

* fix patch

* fix patch
 2025-02-13 23:05:54 +08:00
Billy Cao
11eac71c13 [trainer] fix gen_kwarg to eval during training (#5451) 
* Correctly pass gen_kwarg to eval during model runs

* fix

* fix

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
 2025-02-13 02:35:06 +08:00
SrWYG
1e35967ae1 [data] evaluate on each dataset (#5522) 
* [Update] loader.py , evaluate will run separate evaluations on each dataset.

`If you pass a dictionary with names of datasets as keys and datasets as values, evaluate will run separate evaluations on each dataset. This can be useful to monitor how training affects other datasets or simply to get a more fine-grained evaluation`

seq2seqtrainner support eval_dataset as Dict.

* fix format

* fix

* fix

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
 2025-02-13 02:19:03 +08:00
Noah
4c7bfebcf1 [data] improve error handling (#6128) 
* sync from upstream

* update

* update

* fix

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
 2025-02-13 01:39:41 +08:00
hoshi-hiyouga
617c8ab467 [breaking change] refactor data pipeline (#6901) 
* refactor data

* rename file
 2025-02-13 00:39:20 +08:00
hoshi-hiyouga
e34c3c06da [misc] fix grad ckpt func (#6916) 2025-02-13 00:17:18 +08:00
marko1616
b7fd1e9c00 [trainer] fix llama3.2 vision kto train (#6904) 2025-02-12 19:09:14 +08:00
hoshi-hiyouga
2f8b6847f5 [data] feat: auto template (#6905) 
* support auto template

* add unittest
 2025-02-12 00:22:53 +08:00
hoshi-hiyouga
e1a7c1242c [data] fix ollama template (#6902) 
* fix ollama template

* add meta info

* use half precision
 2025-02-11 22:43:09 +08:00
hoshi-hiyouga
9184a6e0ed [misc] support export ollama modelfile (#6899) 
* support export ollama modelfile

* update config

* add system and num ctx
 2025-02-11 19:52:25 +08:00
hoshi-hiyouga
d1b8aa3835 [data] refactor template (#6896) 2025-02-11 17:59:25 +08:00
hoshi-hiyouga
aca63bfcca [data] refactor mm plugin (#6895) 
* refactor plugin

* lint
 2025-02-11 16:34:49 +08:00
HJ
9153a7bd83 [data] fix qwen_2_5_vl video processing (#6868) 
* fix qwen_2_5_vl video processing

* Update mm_plugin.py

* Update mm_plugin.py

---------

Co-authored-by: hoshi-hiyouga <hiyouga@buaa.edu.cn>
 2025-02-11 16:14:50 +08:00
Zhangchi Feng
764627645a [da'ta] fix minicpmv plugin (#6890) 
* fix template name

* tiny fix

* support minicpm-o-2.6

* support inference of minicpmv

* update readme

* support dpo of minicpmv

* update init audio

* update init audio

* [model]fix image process in minicpmo

* fix no mm inputs
 2025-02-11 13:30:44 +08:00
HJ
0fb44cb3a5 [data] fix: sharegpt converter (#6879) 
* fix-sharegpt-format

* fix

---------

Co-authored-by: hoshi-hiyouga <hiyouga@buaa.edu.cn>
 2025-02-10 21:59:12 +08:00
hoshi-hiyouga
b68199db27 [data] fix mllama collator (#6874) 2025-02-09 22:42:25 +08:00
hoshi-hiyouga
f6f3f8d0fc [test] align test cases (#6865) 
* align test cases

* fix function formatter
 2025-02-09 01:03:49 +08:00
hoshi-hiyouga
1356f9d840 [dataset] add openthought (#6866) 2025-02-09 00:53:01 +08:00
hoshi-hiyouga
5f38bcaba9 [deps] upgrade vllm (#6857) 2025-02-08 15:02:28 +08:00
hoshi-hiyouga
40048ab77a fix qwen2vl plugin (#6855) 2025-02-08 10:59:10 +08:00
hoshi-hiyouga
74ade3a176 [misc] allow extra args (#6831) 2025-02-06 12:38:08 +08:00
Zhangchi Feng
24c7842948 [model] support audio (#6701) 
* support qwen2_audio

* improve code

* lint

* fix

* fix

* fix

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
 2025-02-05 04:59:09 +08:00
Yueqi Song
a5e943f7bc [data] allow thought in function call (#6797) 
* Update template.py

* Update template.py

* use formatter

* fix regex

---------

Co-authored-by: hiyouga <hiyouga@buaa.edu.cn>
 2025-02-05 02:26:23 +08:00
hoshi-hiyouga
e2dc5b952a [misc] update license year & fix llama pro (#6814) 
* fix llamapro script

* change year
 2025-02-05 01:53:33 +08:00
Yueqi Song
dd6b7d203e [data] fix qwen tool template (#6796) 
* Update tool_utils.py

* fix unittest

---------

Co-authored-by: hoshi-hiyouga <hiyouga@buaa.edu.cn>
 2025-02-05 00:02:00 +08:00
Zhangchi Feng
ab9bd068ef [data] fix minicpmv plugin (#6801) 
* fix template name

* tiny fix

* support minicpm-o-2.6

* support inference of minicpmv

* update readme

* support dpo of minicpmv

* update init audio

* update init audio

* [model]fix image process in minicpmo
 2025-02-04 21:20:15 +08:00
hoshi-hiyouga
94803d8133 [model] add mistral small models (#6786) 2025-02-01 04:31:38 +08:00
hoshi-hiyouga
999c7c8fe0 [model] add qwen2.5 vl models (#6779) 2025-01-31 03:00:29 +08:00
hoshi-hiyouga
15357cdad9 [breaking] support transformers 4.48 (#6628) 2025-01-31 01:36:33 +08:00
hoshi-hiyouga
45e68b9f09 [webui] improve webui & reasoning mode (#6778) 2025-01-31 00:09:21 +08:00
qvlehao
28417f862a [model] add deepseek-R1 & show think process (#6767) 2025-01-29 12:16:26 +08:00
yinpu
0f45982bac fix: avoid redundant normalization in DPO's SFT loss calculation (#6722) 2025-01-21 13:38:02 +08:00
engchina
de9bc3fefa [webui] support ja (#6698) 
* add support for japanese language

* add support for japanese language

---------

Co-authored-by: engchina <atjapan2015@gmail.com>
 2025-01-20 19:46:38 +08:00
hoshi-hiyouga
1f47b6186c [model] support yarn (#6693) 2025-01-18 13:56:09 +08:00
hoshi-hiyouga
c0caa7afc6 [misc] update mm plugin (#6691) 2025-01-17 23:04:26 +08:00
hoshi-hiyouga
77bbf65905 disable valset by default (#6690) 2025-01-17 21:09:30 +08:00
hoshi-hiyouga
4d0f662dbe [webui] upgrade to gradio 5 (#6688) 2025-01-17 20:15:42 +08:00
hoshi-hiyouga
7bf09abf1c fix qwen2 moe (#6684) 2025-01-17 13:46:09 +08:00
Zhangchi Feng
027942789b [data] Fix minicpmv/o dpo training (#6657) 
* fix template name

* tiny fix

* support minicpm-o-2.6

* support inference of minicpmv

* update readme

* support dpo of minicpmv
 2025-01-15 17:30:37 +08:00
steveepreston
76675b654e Update val_size english description (#6653) 
* Update `val_size` Description in locales.py

* Update `val_size` Description in data_args.py

* Remove extra space in data_args.py
 2025-01-15 16:00:20 +08:00
hoshi-hiyouga
7a04021d04 [optim] clean apollo (#6645) 
* clean apollo code

* update readme
 2025-01-15 01:42:50 +08:00
zhuHQ
d9189f9f0b [optim] add support to APOLLO (#6617) 2025-01-15 00:24:56 +08:00
hoshi-hiyouga
1278c3e92e lint (#6641) 2025-01-14 18:40:07 +08:00
Haian Huang(深度眸)
deacc00b12 Support InternLM3 Dense 8B Model (#6640) 
* support internlm3

* update

* update

* update

* add hint
 2025-01-14 18:07:27 +08:00
Xiaosu Zhu
58d029f321 Fix tokenizer max length (#6632) 2025-01-14 17:35:54 +08:00
Zhangchi Feng
158a127d34 Support Inference of MiniCPM-V-2.6 and MiniCPM-o-2.6 (#6631) 
* fix template name

* tiny fix

* support minicpm-o-2.6

* support inference of minicpmv
 2025-01-14 17:34:58 +08:00
hoshi-hiyouga
98189c8e4d [model] fix mllama any image (#6637) 
* fix mllama any image

* reorder classes
 2025-01-14 16:47:58 +08:00
hoshi-hiyouga
1c7663d304 pin vllm version to 0.6.5 (#6629) 2025-01-14 02:44:02 +08:00
Zhangchi Feng
c3fda5046d Support new features of MiniCPM-V (#6626) 
* fix template name

* tiny fix

* support minicpm-o-2.6
 2025-01-14 00:26:19 +08:00