Commits

Commits on Feb 10, 2023

update example in README.md
voidful
committedFeb 10, 2023

Commits on Feb 9, 2023

fix encoder decoder model - handle encoder input correctly
voidful
committedFeb 9, 2023
add LICENSE
voidful
committedFeb 9, 2023

Commits on Feb 8, 2023

update remade for key parameter for RL training
voidful
committedFeb 8, 2023

Commits on Feb 6, 2023

enable actor to generate multiple sample for RLHF
voidful
committedFeb 6, 2023

Commits on Feb 2, 2023

enable top K, top P sampling
voidful
committedFeb 2, 2023

Commits on Jan 12, 2023

update 178B training example
voidful
committedJan 12, 2023
Update README.md
voidful
committedJan 12, 2023

Commits on Jan 11, 2023

Merge remote-tracking branch 'origin/main'
voidful
committedJan 11, 2023
add support for petals BLOOM-176B model
voidful
committedJan 11, 2023
Update README.md
voidful
committedJan 11, 2023

Commits on Jan 10, 2023

replace torch.no_grad with inference_mode
voidful
committedJan 10, 2023

Commits on Jan 9, 2023

reduce memory usage by changing optimizer to SGD
voidful
committedJan 9, 2023

Commits on Dec 15, 2022

Update README.md
voidful
committedDec 15, 2022

Commits on Dec 10, 2022

Update README.md
voidful
committedDec 10, 2022

Commits on Dec 9, 2022

stabilize training result & update README.md
voidful
committedDec 9, 2022

Commits on Dec 5, 2022

update pfrl install msg
voidful
committedDec 5, 2022
update pfrl version
voidful
committedDec 5, 2022
re-implement actor and env
voidful
committedDec 5, 2022

Commits on Dec 6, 2021

fix estimate max length incorrect in different pretrained model
voidful
committedDec 6, 2021

Commits on Jul 28, 2021

remove grad on transformer model
voidful
committedJul 28, 2021

Commits on Jul 26, 2021

add dump model & example
voidful
committedJul 26, 2021

Commits on Jul 1, 2021

handle tokenizer and model vocab size mismatch
voidful
committedJul 1, 2021

Commits on Jun 23, 2021

update README.md
voidful
committedJun 23, 2021

Commits on Jun 17, 2021

return prediction
voidful
committedJun 17, 2021
better decoder detection
voidful
committedJun 17, 2021
fix detection may not work on specific model
voidful
committedJun 17, 2021

Commits on Jun 13, 2021

predicted should be empty on different episodes
voidful
committedJun 13, 2021
avoid model exceed max length
voidful
committedJun 13, 2021
fix encoder decoder incorrect input on decoding
voidful
committedJun 13, 2021

Commits on Jun 9, 2021

fix(enc-dec-models-error): hidden state not exist
voidful
committedJun 9, 2021

Commits on Apr 22, 2021

update README.md
voidful
committedApr 22, 2021
feat(update-pypi): uploading to pypi
voidful
committedApr 22, 2021

Commits on Apr 19, 2021

add input text for reward calculation
voidful
committedApr 19, 2021

Commits on Apr 16, 2021

fix(env): retrun dict on env step
voidful
committedApr 16, 2021