Making large AI models cheaper, faster and more accessible
COMMITS
April 28, 2025
Y
fix pp+tp, fix dataloader (#6280)
YeAnbang committed
April 27, 2025
T
fix save issue (#6279)
Tong Li committed
April 26, 2025
Y
fix checkpoint naming; add num_epoch parameter (#6277)
YeAnbang committed
April 25, 2025
Y
[feat] Support DAPO (#6263)
YeAnbang committed
April 23, 2025
T
[feat] Add final save at the end (#6274)
Tong Li committed
April 22, 2025
T
add prompt template (#6273)
Tong Li committed
April 10, 2025
Y
[hot-fix] Fix memory leakage bug, support TP+PP (#6258)
YeAnbang committed
April 9, 2025
Y
[Distributed RLHF] Integration of PP (#6257)
YeAnbang committed
March 28, 2025
Y
[feat] add microbatch forwarding (#6251)
YeAnbang committed
March 21, 2025
Y
update logging
YeAnbang committed
Y
simplify vllm preprocessing input ids
YeAnbang committed
Y
fix logprob, add filtering, temperature annealing, lr descent
YeAnbang committed
March 18, 2025
D
[Feature] Support Distributed LogProb for GRPO Training (#6247)
duanjunwen committed
March 14, 2025
Y
fix transformers backend
YeAnbang committed
March 13, 2025
T
setup update
Tong Li committed
T
print results
Tong Li committed
T
convert to 8 generation
Tong Li committed
T
fix consumer
Tong Li committed
T
fix tp bug
Tong Li committed
March 11, 2025
T
add response length
Tong Li committed
T
fix reward score
Tong Li committed
March 10, 2025
T
update reward
Tong Li committed
T
update reward fn
Tong Li committed
T
update grpo
Tong Li committed
March 6, 2025
P
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] committed
P
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] committed
T
update select algo
Tong Li committed