added wandb in train file a0e028b
Bhaskar commited on
Commit History
Fix Unsloth dtype mismatch by locking precision mode 91ac1b6
Bhaskar commited on
Fix grpo reward parser for list/dist completions 8ca88e8
Bhaskar commited on
add colab notebook and updated the train and docker files and edited the requirements according to the codebase 2b9c8cb
Bhaskar commited on
updated the codebase for a10 24gb gpu a8a8219
Bhaskar commited on
updated the codebase for a10 24gb gpu a24d3c8
Bhaskar commited on
train file updated for local training 4714235
Bhaskar commited on
Round 2 Upgrade: Added GRPO train.py and vector-field reward shaping 51bb0d4
Bhaskar commited on