<< Back to research
fw-ai/cookbook/researchskill versions
4 skill versions publishedCompare skill versions
v1.0.3currentAutomated scan4h ago
[codex] make GLM5 preserve-thinking SFT default (#397)
sha256:1aea96ea8fff
v1.0.2Automated scan100/1001d ago
Rename PPO snapshot logprobs (#396)
sha256:f62cce242db9
v1.0.1Automated scan100/1001d ago
feat(checkpoints): add TrainingCheckpoints with resumable/promotable axes (#385)
sha256:a935c1215f7e
v1.0.0LLM-judged scan90/1002d ago
sha256:d5c045513c45