Hugging Face's logo

shivam2k3
/
opensoc-env

openenv
cybersecurity
rlvr
self-play
Model card Files
xet
 Community
opensoc-env / train /sft_warmstart.py

Commit History

sft_warmstart: import unsloth first; batched formatting_func
99d0d29

shivam2k3 commited on

sft_warmstart: hardcode Qwen2.5 eos_token <|im_end|>
b42f9bf

shivam2k3 commited on

sft_warmstart: replace unsloth's <EOS_TOKEN> placeholder with <|im_end|>
67d026b

shivam2k3 commited on

sft_warmstart: pass real eos_token (trl 0.24 sentinel bug)
c3488dd

shivam2k3 commited on

sft_warmstart: trl 0.24 API (SFTConfig + processing_class)
a5f5c45

shivam2k3 commited on

OpenSOC v1
bb6a031

shivam2k3 commited on