GitLab

Optimization Code
I think there were two extra $ entered here, which have been deleted

* run the base

* working on dist ppo

* sync

* detached trainer

* update detached trainer. no maker update function

* facing init problem

* 1 maker 1 trainer detached run. but no model update

* facing cuda problem

* fix save functions

* verified maker update

* nothing

* add ignore

* analyize loss issue

* remove some debug codes

* facing 2m1t stuck issue

* 2m1t verified

* do not use torchrun

* working on 2m2t

* working on 2m2t

* initialize strategy in ray actor env

* facing actor's init order issue

* facing ddp model update issue (need unwarp ddp)

* unwrap ddp actor

* checking 1m2t stuck problem

* nothing

* set timeout for trainer choosing. It solves the stuck problem!

* delete some debug output

* rename to sync with upstream

* rename to sync with upstream

* coati rename

* nothing

* I am going to detach the replaybuffer from trainer and make it a Ray Actor. Two benefits: 1. support TP trainer. 2. asynchronized buffer operations

* experience_maker_holder performs target-revolving _send_experience() instead of length comparison.

* move code to ray subfolder

* working on pipeline inference

* apply comments

---------
Co-authored-by: csric <richcsr256@gmail.com>

Display format optimization , same as fix#3562
Simultaneous modification of en version

* [misc] add print verbose

* [gemini] add print verbose

* [zero] add print verbose for low level

* [misc] add print verbose for op builder

Display format optimization, fix bug#3562
Specific changes
1. "This is called a column-parallel fashion" Translate to Chinese
2. use the ```math code block syntax to display a math expression as a block, No modification of formula content

Please check that the math formula is displayed correctly
If OK, I will change the format of the English version of the formula in parallel

* [feat][chatgpt]train prompts on ray example

* [fix]simplify code

* [fix]remove depreciated parameter

* [fix]add dependencies

* [fix]method calling

* [fix]experience maker

* [fix]missing loss function

* [fix]init optimizer

* [feat]add usage comment

* [fix]rename files

* [fix]add readme

* [fix]file path

* [fix]move directory

---------
Co-authored-by: jiangwen <zxl265370@antgroup.com>

* [chat] clean up duplicate tutorial

* [chat] clean up duplicate tutorial

* [chat] clean up duplicate tutorial

* [chat] clean up duplicate tutorial

Format Optimization ,Add [] outside of DeepSpeed

Delete more ")"

Fixing document link errors using absolute paths

* [gemini] fix nvme optimizer init

* [gemini] gemini supports lazy init

* [gemini] add init example

* [gemini] add fool model

* [zero] update gemini ddp

* [zero] update init example

* add chunk method

* add chunk method

* [lazyinit] fix lazy tensor tolist

* [gemini] fix buffer materialization

* [misc] remove useless file

* [booster] update gemini plugin

* [test] update gemini plugin test

* [test] fix gemini plugin test

* [gemini] fix import

* [gemini] fix import

* [lazyinit] use new metatensor

* [lazyinit] use new metatensor

* [lazyinit] fix __set__ method

[checkpoint]  Shard saved checkpoint need to be compatible with the naming format of hf checkpoint files  (#3479)

* [checkpoint] support huggingface style sharded checkpoint, to be compatible with hf file naming format

* [checkpoint] support huggingface style sharded checkpoint, to be compatible with hf file naming format

* [checkpoint] Shard saved checkpoint add 'variant' field to customize filename

* [checkpoint] Shard saved checkpoint add 'variant' field to customize filename

* [checkpoint] Shard saved checkpoint add 'variant' field to customize filename

* [checkpoint] Shard saved checkpoint add 'variant' field to customize filename

---------
Co-authored-by: luchen <luchen@luchendeMacBook-Pro.local>
Co-authored-by: luchen <luchen@luchendeMBP.lan>

Co-authored-by: Yuanchen Xu <yuanchen.xu00@gmail.com>

- [ ] Stable Diffusion
- [ ] Dreambooth
It's easy for users to think that we don't support them yet. Add them after migrating them from example to application
https://github.com/hpcaitech/ColossalAI/tree/main/examples/images

* [doc] add requirement and highlight application

* [doc] link example and application

* Update README.md

* Update README.md

* Update README.md

* Update README.md

---------
Co-authored-by: Fazzie-Maqianli <55798671+Fazziekey@users.noreply.github.com>

* mv LlamaForCausalLM to LlamaModel

* rm unused imports

---------
Co-authored-by: gongenlei <gongenlei@baidu.com>

* update roberta example

* update roberta example

* modify conflict & update roberta

* update roberta example

* update roberta example

* Update requirements.txt

* Update environment.yaml

* Update README.md

* Update environment.yaml

* Update README.md

* Update README.md

* Delete requirements_colossalai.txt

* Update requirements.txt

* Update README.md

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

* [checkpoint] support huggingface style sharded checkpoint

---------
Co-authored-by: luchen <luchen@luchendeMBP.lan>

* [test] added spawn decorator

* polish code

* polish code

* polish code

* polish code

* polish code

* polish code

* Update ppo.py

Fix the bug of fetching wrong batch data

* Add peft model support in SFT and Prompts training

In stage-1 and stage-3, the peft model supports are added. So the trained artifacts will be only a small lora additions instead of the whole bunch of files.

* Delete test_prompts.txt

* Delete test_pretrained.txt

* Move the peft stuffs to a community folder.

* Move the demo sft to community

* delete dirty files

* Add instructions to install peft using source

* Remove Chinese comments

* remove the Chinese comments

The function save_model should be a part of PPOTrainer.

* fix stage 2

fix stage 2

* add torch