Mirrors/modelscope
1
0
Fork 0
mirror of https://github.com/modelscope/modelscope.git synced 2025-12-24 03:59:23 +01:00
Code Issues Packages Projects Releases Wiki Activity
1,937 Commits 263 Branches 106 Tags
43a57fe110db42637e7013f6a3d18f04d74fd085
Commit Graph

86 Commits

Author SHA1 Message Date
zsl01670416
43a57fe110 support load dataset for llama 
support loading dataset for llama:
1.load dataset by MsDataset when parameters train dataset name and val dataset name were set. but there is no suitable dataset in hub.
2.load dataset by MsDataset when only parameter train dataset name was set, and then split into train dataset and validation dataset . 
3.load dataset by MsDataset when user set parameter src_txt, which is a file path such as 'alpaca_data.json', and then split into training dataset and validation dataset.
4.load dataset by build dataset from file in flex training.

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13505335
 2023-08-07 19:48:36 +08:00
lukeming.lkm
bd2f70a6eb add quantization in qwen pipelines and relevant unittests 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13499600

* add quant features

* resolve import

* resolve format

* fix save vocab
 2023-08-02 14:05:13 +08:00
lukeming.lkm
33bd74a7be add qwen 7b base and chat 
添加QWen 7b base模型和chat模型及相关pipelines
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13482235

* add qwen 7b base and chat

* fix logger

* update examples, lint test

* add unittest for qwen base and chat

* rename qwen to qwen-7b

* resolve imports and add a registry to text-generation

* reset load model from pretrained

* fix precheck

* skip qwen test case now

* remove strange file
 2023-08-02 09:25:21 +08:00
wenmeng.zwm
48a39244f1 Merge branch master-merge-github-230728 into master 
Title: Merge branch 'master-github' into master-merge-github-230728 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13456249
 2023-07-31 16:09:07 +08:00
suluyan.sly
d7cd2ce28e Merge branch 'master-github' into master-merge-github-230728 2023-07-29 10:42:54 +08:00
Jintao
312b63fe06 fix checkpoint, same device bug (#427) 2023-07-29 00:06:27 +08:00
zsl01670416
847607ab66 Merge branch 'debug_chatglm6b_json_dataset' fix conflict between hf dataset and to_hf_dataset 
The type of dataset built from file is hf dataset, which can not use function to_hf_dataset.
 2023-07-28 19:12:21 +08:00
suluyan.sly
05e1357c32 Merge branch 'master-github' into master-merge-github-230728 2023-07-28 16:40:34 +08:00
Wang Qiang
34ea2b474a Upgrade stable diffusion version to a more powerful version 2.1 (#415) 2023-07-27 20:40:38 +08:00
suluyan.sly
1c6f5fe775 Merge branch 'master-github' into master-merge-github-230727 
Conflicts:
       examples/pytorch/baichuan/finetune_baichuan.py
       examples/pytorch/chatglm6b/finetune.py
 2023-07-27 17:29:27 +08:00
Wang Qiang
66cf72a75c Merge pull request #376 from XDUWQ/custom_diffusion 
Custom method for finetuning stable diffusion
 2023-07-27 10:41:38 +08:00
Jintao
4ca937d2ba support openbuddy-llama2-13b (#416) 2023-07-26 18:12:55 +08:00
XDUWQ
99aa707995 fix bugs 2023-07-26 16:35:40 +08:00
tastelikefeet
0db3d1d53b Fix bug of amp and device_map (#397) 
* fix amp

* remove useless code

* Fix bug
 2023-07-25 19:28:00 +08:00
Jintao
f03898626e ckpt output directory ignore *.safetensors (#410) 
* ckpt output file ignore *.safetensors

* update
 2023-07-25 19:27:11 +08:00
hemu.zp
fc54593a56 fix baichuan eval and support sequence_length 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13404289
* fix baichuan eval

* support sequence_length and ppl

* fix typo

* fix bug for palm

* fix bug
 2023-07-25 19:10:45 +08:00
zsl01670416
9926ad685b support getting labels from dataset in sbert text classification and building dataset from file in chatglm-6b 
1.Add getting labels from dataset in "text_classificationfinetune_text_classification.py" to simplify user's operation in flex training. Parameters "--num_labels" and "--labels" were removed in "run_train.sh".
2.In "chatglm6b / finetune.py", building dataset from file  is necessary to support flex training.
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13382745
* support getting labels from dataset in sbert text classification and building dataset from file in chatglm-6b

* support getting labels from dataset in sbert text classification and building dataset from file in chatglm-6b

* remove repetitive labels in a concise manner of using set

* reserve parameter labels in finetune_text_classification

* Merge branch 'master' of http://gitlab.alibaba-inc.com/Ali-MaaS/MaaS-lib

reserve parameter labels in finetune_text_classification

* Merge branch 'support_text_cls_labels_chatglm_json'
reserve parameter labels in finetune_text_classification
 2023-07-25 19:02:32 +08:00
hemu.zp
ed6e139759 Support llama & lora finetune without deepspeed 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13131145
* support llama + lora without deepspeed

* feat: Fix conflict, auto commit by WebIDE
 2023-07-25 17:32:46 +08:00
XDUWQ
3412a074c5 precommit 2023-07-25 15:00:28 +08:00
XDUWQ
ffbf77fcf2 update 2023-07-25 14:47:45 +08:00
XDUWQ
8e157cfa15 precommit 2023-07-24 22:10:52 +08:00
XDUWQ
426f55d57b add lora_rank for lora stable diffusion 2023-07-24 19:43:20 +08:00
XDUWQ
bc93e2dc96 add lora_rank for lora stable diffusion 2023-07-24 19:32:04 +08:00
XDUWQ
eb24e23d19 add lora_rank for lora stable diffusion 2023-07-24 19:24:52 +08:00
XDUWQ
6fb340e7f8 add lora_rank for lora stable diffusion 2023-07-24 19:17:49 +08:00
Jintao
ba4b9fc43f Added full parameter sft to llm (#402) 
* Optimized code

* update parse_args

* fix get_logger bug

* update parse_args

* Added full parameter fine-tuning

* Add support_bf16 warning

* Modify the code format and fix bugs
 2023-07-24 15:52:09 +08:00
xingjun.wang
96a5282021 check format 2023-07-21 23:09:24 +08:00
Jintao
2f7c669f33 support llama2 (#393) 
* Unify sft and infer code into a single file

* update llama2 sft infer
 2023-07-19 17:34:27 +08:00
tastelikefeet
12bc1603a9 Fix amp + device_map (#386) 
1. Fix the amp + device_map bug in chatglm2 finetune code 
2. Optional to save optimizer state 
3. Fix the logger double print problem
 2023-07-16 08:45:20 +08:00
Jintao
c6189d68a0 Fix/chatglm2 (#384) 2023-07-15 09:59:53 +08:00
XDUWQ
bcf443c672 custom diffusion 2023-07-12 15:32:48 +08:00
XDUWQ
0e99c27d54 custom diffusion 2023-07-12 15:27:08 +08:00
XDUWQ
d6368b2617 custom diffusion 2023-07-12 15:02:43 +08:00
XDUWQ
0dc57f8dec custom diffusion 2023-07-12 14:56:14 +08:00
Xingjun.Wang
0f36b081ef Merge pull request #371 from foocker/master 
ASRDataset for download_mode parameters
 2023-07-12 14:47:32 +08:00
tastelikefeet
544f6c0410 Fea/chatglm6b v2 new version (#368) 
* upgrade code

* add chatglm2 ptuning

* pre-commit passed
 2023-07-12 14:44:09 +08:00
Wang Qiang
04b940f600 Merge branch 'modelscope:master' into custom_diffusion 2023-07-12 10:08:11 +08:00
gg
49c6d8bcf6 pre-commit 2023-07-12 09:43:24 +08:00
XDUWQ
1caa45422c custom diffusion 2023-07-11 20:46:32 +08:00
gg
574b4568ff flake8 2023-07-11 18:36:54 +08:00
Jintao
d20d033e07 add example/llm (#372) 
* add example/llm

* fix lint test
 2023-07-11 17:35:11 +08:00
fq
d47684de5b Optimize comments 2023-07-11 17:12:59 +08:00
fq
4d77e57769 add download_mode param to params 
maybe set from funasr is better.
 2023-07-11 16:43:37 +08:00
fq
61faedfb15 Update finetune_speech_recognition.py add params.download_mode 
set from params, config
 2023-07-11 15:18:59 +08:00
fq
2073f4fd55 Update finetune_speech_recognition.py 
MsDataset replace by ASRDataset.
 2023-07-11 11:54:05 +08:00
fq
20c15d3aaa Update finetune_speech_recognition.py 
using the newest ASRDataset, and add download_mode for re-download the dataset(dataset is broken and so on)
 2023-07-11 11:33:18 +08:00
Wang Qiang
d49953b943 fix bugs of loading local sd dataset (#357) 2023-07-04 22:01:21 +08:00
Firmament-cyou
423e2ce940 Add lora_inference for baichuan. (#352) 
* add lora_inference.py for baichuan

* fix linttest

* fix linttest

---------

Co-authored-by: hemu <hemu.zp@alibaba-inc.com>
 2023-07-04 18:39:36 +08:00
tastelikefeet
08c71f1f3d Fix/chatglm6b 2 (#354) 2023-07-04 01:58:57 +08:00
tastelikefeet
45cf0035f4 fix chatglm2 evaluation error: hypothesis emtpy (#348) 
* fix evaluation error: hypothesis emtpy

* fix pipeline

* fix bug
 2023-07-03 23:16:38 +08:00