GitLab

* [chat] fix gemini strategy

* [chat] fix gemini strategy

* [chat] fix gemini strategy

* [chat] fix gemini strategy

* g# This is a combination of 2 commits.

[chat] fix gemini strategy

fox

* [chat] fix gemini strategy

update llama2 example

[chat] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* [fix] fix gemini strategy

* fix

* fix

* fix

* fix

* fix

* Update train_prompts.py

* feat: modify forward fn of critic and reward model

* feat: modify calc_action_log_probs

* to: add wandb in sft and rm trainer

* feat: update train_sft

* feat: update train_rm

* style: modify type annotation and add warning

* feat: pass tokenizer to ppo trainer

* to: modify trainer base and maker base

* feat: add wandb in ppo trainer

* feat: pass tokenizer to generate

* test: update generate fn tests

* test: update train tests

* fix: remove action_mask

* feat: remove unused code

* fix: fix wrong ignore_index

* fix: fix mock tokenizer

* chore: update requirements

* revert: modify make_experience

* fix: fix inference

* fix: add padding side

* style: modify _on_learn_batch_end

* test: use mock tokenizer

* fix: use bf16 to avoid overflow

* fix: fix workflow

* [chat] fix gemini strategy

* [chat] fix

* sync: update colossalai strategy

* fix: fix args and model dtype

* fix: fix checkpoint test

* fix: fix requirements

* fix: fix missing import and wrong arg

* fix: temporarily skip gemini test in stage 3

* style: apply pre-commit

* fix: temporarily skip gemini test in stage 1&2

---------
Co-authored-by: Mingyan Jiang <1829166702@qq.com>

* [misc] update pre-commit

* [misc] run pre-commit

* [misc] remove useless configuration files

* [misc] ignore cuda for clang-format

* style: rename replay buffer

Experience replay is typically for off policy algorithms.
Use this name in PPO maybe misleading.

* fix: fix wrong zero2 default arg

* test: update experience tests

* style: rename zero_pad fn

* fix: defer init in CycledDataLoader

* test: add benchmark test

* style: rename internal fn of generation

* style: rename internal fn of lora

* fix: remove unused loss fn

* fix: remove unused utils fn

* refactor: remove generate_with_actor fn

* fix: fix type annotation

* test: add models tests

* fix: skip llama due to long execution time

* style: modify dataset

* style: apply formatter

* perf: update reward dataset

* fix: fix wrong IGNORE_INDEX in sft dataset

* fix: remove DataCollatorForSupervisedDataset

* test: add dataset tests

* style: apply formatter

* style: rename test_ci to test_train

* feat: add llama in inference

* test: add inference tests

* test: change test scripts directory

* fix: update ci

* fix: fix typo

* fix: skip llama due to oom

* fix: fix file mod

* style: apply formatter

* refactor: remove duplicated llama_gptq

* style: apply formatter

* to: update rm test

* feat: add tokenizer arg

* feat: add download model script

* test: update train tests

* fix: modify gemini load and save pretrained

* test: update checkpoint io test

* to: modify nproc_per_node

* fix: do not remove existing dir

* fix: modify save path

* test: add random choice

* fix: fix sft path

* fix: enlarge nproc_per_node to avoid oom

* fix: add num_retry

* fix: make lora config of rm and critic consistent

* fix: add warning about lora weights

* fix: skip some gpt2 tests

* fix: remove grad ckpt in rm and critic due to errors

* refactor: directly use Actor in train_sft

* test: add more arguments

* fix: disable grad ckpt when using lora

* fix: fix save_pretrained and related tests

* test: enable zero2 tests

* revert: remove useless fn

* style: polish code

* test: modify test args

* feat: remove on_learn_epoch fn as not used

* revert: add _on_learn_epoch fn

* to: remove the use of NaiveStrategy

* test: remove NaiveStrategy tests

* feat: remove NaiveStrategy

* style: modify comments and params

* feat: split ColossalAIStrategy into LowLevelZeroStrategy and GeminiStrategy

* fix: remove naive

* fix: align with modified colossal strategy

* fix: fix ddp _try_init_dist arg

* to: add SLTrainer

* refactor: refactor RMTrainer and SFTTrainer

* fix: fix init file

* feat: remove on_learn_epoch fn as not used

* fix: align with modified gemini arguments

* to: add OnPolicyTrainer

* revert: add _on_learn_epoch fn

* refactor: refactor PPOTrainer

* style: rename PPOTrainer argument

* fix: align with modified PPO arguments

* test: align with modified train_prompts arguments

* chore: modify train_prompts

* docs: align with modified arguments

* fix: remove unnecessary output

* fix: move dataloader to fit fn of SLTrainer

* fix: move dataloader to fit fn of OnPolicyTrainer

* fix: modify usage of prompt and pretrain dataloader

* refactor: adapt boost API in base and naive strategies

* fix: initialize plugin after setup_distributed

* fix: fix save_pretrained fn

* refactor: adapt boost API in DDPStrategy

* to: add _post_init check

* to: fix ddp backward, modify ddp dataloader and unwrap

* feat: adapt boost API in ColossalAIStrategy

* fix: call setup_distributed before use get_current_device

* fix: fix save_model and save_optimizer

* test: remove save_sharded_optimizer test

* style: apply formatter

* fix: fix stage check and add comments

* feat: allow dict type arg in strategy.prepare

* to: temporarily remove lr_scheduler for testing

* style: simplify init of ColossalAIStrategy

* fix: fix lr_scheduler in sft and rm

* style: modify comments

* test: add train_prompts tests

* fix: fix inference only case and use in train_prompts

* test: skip failed tests in ci

* style: fix CodeFactor check

* fix: do not use model.to('cpu') with GeminiPlugin

* test: enable colossalai_gemini tests

* test: set CUDA_VISIBLE_DEVICES in ci

* docs: add note

* refactor: separate log_probs fn from Actor forward fn

* refactor: separate generate fn from Actor class

* feat: update unwrap_model and get_base_model
* unwrap_model returns model not wrapped by Strategy
* get_base_model returns HF model for Actor, Critic and RewardModel

* feat: simplify Strategy.prepare

* style: remove get_base_model method of Actor

* perf: tokenize text in batches

* refactor: move calc_action_log_probs to utils of model

* test: update test with new forward fn

* style: rename forward fn args

* fix: do not unwrap model in save_model fn of naive strategy

* test: add gemini test for train_prompts

* fix: fix _set_default_generate_kwargs

* Detached ppo (#9)

* 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

* working on pipeline strategy. in progress.

* remove pipeline code. clean this branch

* update remote parameters by state_dict. no test

* nothing

* state_dict sharding transfer

* merge debug branch

* gemini _unwrap_model fix

* simplify code

* simplify code & fix LoRALinear AttributeError

* critic unwrapped state_dict

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

* [chat] add perfomance evaluator and fix bugs (#10)

* [chat] add performance evaluator for ray

* [chat] refactor debug arg

* [chat] support hf config

* [chat] fix generation

* [chat] add 1mmt dummy example

* [chat] fix gemini ckpt

* split experience to send (#11)
Co-authored-by: csric <richcsr256@gmail.com>

* [chat] refactor trainer and maker (#12)

* [chat] refactor experience maker holder

* [chat] refactor model init

* [chat] refactor trainer args

* [chat] refactor model init

* [chat] refactor trainer

* [chat] refactor experience sending logic and training loop args (#13)

* [chat] refactor experience send logic

* [chat] refactor trainer

* [chat] refactor trainer

* [chat] refactor experience maker

* [chat] refactor pbar

* [chat] refactor example folder (#14)

* [chat] support quant (#15)

* [chat] add quant

* [chat] add quant example

* prompt example (#16)

* prompt example

* prompt load csv data

* remove legacy try

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

* [chat] add mmmt dummy example and refactor experience sending (#17)

* [chat] add mmmt dummy example

* [chat] refactor naive strategy

* [chat] fix struck problem

* [chat] fix naive strategy

* [chat] optimize experience maker sending logic

* [chat] refactor sending assignment

* [chat] refactor performance evaluator (#18)

* Prompt Example & requires_grad state_dict & sharding state_dict (#19)

* prompt example

* prompt load csv data

* remove legacy try

* maker models require_grad set to False

* working on zero redundancy update

* mmmt_prompt example; naive strategy requires_grad state_dict & sharding; maker model requires_no_grad.

* remove legacy examples

* remove legacy examples

* remove replay buffer tp state. bad design

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

* state_dict sending adapts to new unwrap function (#20)

* prompt example

* prompt load csv data

* remove legacy try

* maker models require_grad set to False

* working on zero redundancy update

* mmmt_prompt example; naive strategy requires_grad state_dict & sharding; maker model requires_no_grad.

* remove legacy examples

* remove legacy examples

* remove replay buffer tp state. bad design

* opt benchmark

* better script

* nothing

* [chat] strategy refactor unwrap model

* [chat] strategy refactor save model

* [chat] add docstr

* [chat] refactor trainer save model

* [chat] fix strategy typing

* [chat] refactor trainer save model

* [chat] update readme

* [chat] fix unit test

* working on lora reconstruction

* state_dict sending adapts to new unwrap function

* remove comments

---------
Co-authored-by: csric <richcsr256@gmail.com>
Co-authored-by: ver217 <lhx0217@gmail.com>

* [chat-ray] add readme (#21)

* add readme

* transparent graph

* add note background

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

* [chat] get images from url (#22)

* Refactor/chat ray (#23)

* [chat] lora add todo

* [chat] remove unused pipeline strategy

* [chat] refactor example structure

* [chat] setup ci for ray

* [chat-ray] Support LoRA trainer. LoRA weights reconstruction. (#24)

* lora support prototype

* lora support

* 1mmt lora & remove useless code

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

* [chat] fix test ci for ray

* [chat] fix test ci requirements for ray

* [chat] fix ray runtime env

* [chat] fix ray runtime env

* [chat] fix example ci docker args

* [chat] add debug info in trainer

* [chat] add nccl debug info

* [chat] skip ray test

* [doc] fix typo

---------
Co-authored-by: csric <59389055+CsRic@users.noreply.github.com>
Co-authored-by: csric <richcsr256@gmail.com>

* [chat] strategy refactor unwrap model

* [chat] strategy refactor save model

* [chat] add docstr

* [chat] refactor trainer save model

* [chat] fix strategy typing

* [chat] refactor trainer save model

* [chat] update readme

* [chat] fix unit test

* [chat] refactor lora

* [chat] remove lm class

* [chat] refactor save model

* [chat] refactor train sft

* [chat] fix ci

* [chat] fix ci

* [gemini] support don't scatter after inference

* [chat] update colossalai strategy

* [chat] fix opt benchmark

* [chat] update opt benchmark

* [gemini] optimize inference

* [test] add gemini inference test

* [chat] fix unit test ci

* [chat] fix ci

* [chat] fix ci

* [chat] skip checkpoint test

* [zero] refactor low-level zero folder structure

* [zero] fix legacy zero import path

* [zero] fix legacy zero import path

* [zero] remove useless import

* [zero] refactor gemini folder structure

* [zero] refactor gemini folder structure

* [zero] refactor legacy zero import path

* [zero] refactor gemini folder structure

* [zero] refactor gemini folder structure

* [zero] refactor gemini folder structure

* [zero] refactor legacy zero import path

* [zero] fix test import path

* [zero] fix test

* [zero] fix circular import

* [zero] update import

* fix colo-stratergy

* polish

* fix lora

* fix ddp

* polish

* polish

* [chatgpt] startegy add prepare method

* [chatgpt] refactor examples

* [chatgpt] refactor strategy.prepare

* [chatgpt] support save/load checkpoint

* [chatgpt] fix unwrap actor

* [chatgpt] fix unwrap actor