23 days ago
-
Final agent GRPO main; mean_reward=0.100
-
TRL: inject text/JSON tool calls + forced ls into tool loop
-
Force first tool + tool_choice required nudge
-
Shaped rewards + root listing in reset
-
Fix Langfuse ckpt/heldout child trace_context
-
Fix Langfuse: clean roots + explicit child parents
-
Ultra-detailed agent prompt
-
Reset tip edge cases
-
Smoke mode + detailed train entry
-
Add deepagents_rl package sources
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
ls fallback: no assume src/
-
ls fallback: no assume src/
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add kaggle/train_grpo.py
-
Add kaggle/train_grpo.py
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add kaggle/KAGGLE_SETUP.md
-
Add kaggle/train_grpo.py
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add README.md
-
Add configs/kaggle_grpo.yaml
-
Add kaggle/KAGGLE_SETUP.md
-
Add kaggle/train_grpo.py
-
Add kaggle/KAGGLE_SETUP.md
-
Add kaggle/train_grpo.py
-
Add kaggle/KAGGLE_SETUP.md
-
Add kaggle/train_grpo.py
-
Add kaggle/train_grpo.py
-
Add deepagents_rl package sources
-
Add README.md
-
Add configs/kaggle_grpo.yaml
-
Add kaggle/KAGGLE_SETUP.md
-
Add kaggle/train_grpo.py
-
initial commit