Oct 22 22 days ago

1. 13:14

   Pass train_dataset (required by AsyncGRPOTrainer ctor in 1.15.0)

   SaylorTwift HF Staff
   Verified signature

   2a4d2a5
2. 13:10

   Add kernels==0.16.0 (trl 1.15.0 loads kernels-community/flash-attn3)

   SaylorTwift HF Staff
   Verified signature

   8067c88
3. 13:05

   Pin trl==1.15.0 (pre-tree-packing, LFM2 hybrid compatible) + vllm v0.31.0 image

   SaylorTwift HF Staff
   Verified signature

   73712c7
4. 12:53

   Install git in job (vllm image lacks it)

   SaylorTwift HF Staff
   Verified signature

   9cd9f57
5. 12:52

   Use python3 (vllm image has no python alias)

   SaylorTwift HF Staff
   Verified signature

   ea9b383
6. 12:50

   Switch sandboxes to Harbor hf-sandbox backend; add pinned task prep with docker_image patch; dataset discovery via prompt_rows()

   SaylorTwift HF Staff
   Verified signature

   a7525f2
7. 12:42

   OpenCode-only LoRA RL on LFM2.5-2.6B (verified against trl 1.16.0 / openenv 0.10.0)

   SaylorTwift HF Staff
   Verified signature

   f0273c6
8. 12:42

   initial commit

   SaylorTwift HF Staff
   Verified signature

   2b008b7