sanjuhs commited on
Commit
fd29c31
·
verified ·
1 Parent(s): 095b2cc

Add strict GRPO report artifacts

Browse files
.gitattributes CHANGED
@@ -105,3 +105,7 @@ data/reference-metrics/shaft_torque_clamp_120nm/glb_reference/glb_reference_norm
105
  data/reference-metrics/shelf_bracket_triangular_200n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
106
  data/reference-metrics/table_six_leg_500n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
107
  data/reference-metrics/wall_j_hook_120n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
105
  data/reference-metrics/shelf_bracket_triangular_200n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
106
  data/reference-metrics/table_six_leg_500n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
107
  data/reference-metrics/wall_j_hook_120n/glb_reference/glb_reference_normalized.stl filter=lfs diff=lfs merge=lfs -text
108
+ training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_code_health.png filter=lfs diff=lfs merge=lfs -text
109
+ training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_reward_curve.png filter=lfs diff=lfs merge=lfs -text
110
+ training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/optimizer_health.png filter=lfs diff=lfs merge=lfs -text
111
+ training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/sft_loss_curve.png filter=lfs diff=lfs merge=lfs -text
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/debug_completion_summary.json ADDED
@@ -0,0 +1,3842 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "reward_call": 1,
4
+ "index": 0,
5
+ "score": -0.84,
6
+ "task_id": "bottle_cage_bike",
7
+ "code_chars": 6180,
8
+ "cadforge_build": 0.0,
9
+ "cadforge_total": -1.0,
10
+ "has_fixture": false,
11
+ "has_cadquery_import": true,
12
+ "error_kind": "SyntaxError"
13
+ },
14
+ {
15
+ "reward_call": 1,
16
+ "index": 1,
17
+ "score": -0.6649999999999999,
18
+ "task_id": "bottle_cage_bike",
19
+ "code_chars": 5944,
20
+ "cadforge_build": 0.0,
21
+ "cadforge_total": -1.0,
22
+ "has_fixture": true,
23
+ "has_cadquery_import": true,
24
+ "error_kind": "SyntaxError"
25
+ },
26
+ {
27
+ "reward_call": 1,
28
+ "index": 2,
29
+ "score": -0.565,
30
+ "task_id": "pipe_clamp_saddle",
31
+ "code_chars": 4746,
32
+ "cadforge_build": 0.0,
33
+ "cadforge_total": -1.0,
34
+ "has_fixture": true,
35
+ "has_cadquery_import": true,
36
+ "error_kind": "TypeError"
37
+ },
38
+ {
39
+ "reward_call": 1,
40
+ "index": 3,
41
+ "score": 0.770444555010733,
42
+ "task_id": "pipe_clamp_saddle",
43
+ "code_chars": 6153,
44
+ "cadforge_build": 1.0,
45
+ "cadforge_total": 0.7299347706008623,
46
+ "has_fixture": true,
47
+ "has_cadquery_import": true,
48
+ "error_kind": "build_ok"
49
+ },
50
+ {
51
+ "reward_call": 2,
52
+ "index": 0,
53
+ "score": 0.8845503697602638,
54
+ "task_id": "table_six_leg_500n",
55
+ "code_chars": 5131,
56
+ "cadforge_build": 1.0,
57
+ "cadforge_total": 0.8641769056003104,
58
+ "has_fixture": true,
59
+ "has_cadquery_import": true,
60
+ "error_kind": "build_ok"
61
+ },
62
+ {
63
+ "reward_call": 2,
64
+ "index": 1,
65
+ "score": -0.565,
66
+ "task_id": "table_six_leg_500n",
67
+ "code_chars": 6073,
68
+ "cadforge_build": 0.0,
69
+ "cadforge_total": -1.0,
70
+ "has_fixture": true,
71
+ "has_cadquery_import": true,
72
+ "error_kind": "unknown"
73
+ },
74
+ {
75
+ "reward_call": 2,
76
+ "index": 2,
77
+ "score": -0.695,
78
+ "task_id": "robot_gripper_two_finger",
79
+ "code_chars": 4208,
80
+ "cadforge_build": 0.0,
81
+ "cadforge_total": -1.0,
82
+ "has_fixture": true,
83
+ "has_cadquery_import": true,
84
+ "error_kind": "SyntaxError"
85
+ },
86
+ {
87
+ "reward_call": 2,
88
+ "index": 3,
89
+ "score": -0.6649999999999999,
90
+ "task_id": "robot_gripper_two_finger",
91
+ "code_chars": 6186,
92
+ "cadforge_build": 0.0,
93
+ "cadforge_total": -1.0,
94
+ "has_fixture": true,
95
+ "has_cadquery_import": true,
96
+ "error_kind": "SyntaxError"
97
+ },
98
+ {
99
+ "reward_call": 3,
100
+ "index": 0,
101
+ "score": -0.84,
102
+ "task_id": "bike_accessory_mount_120n",
103
+ "code_chars": 6128,
104
+ "cadforge_build": 0.0,
105
+ "cadforge_total": -1.0,
106
+ "has_fixture": true,
107
+ "has_cadquery_import": true,
108
+ "error_kind": "SyntaxError"
109
+ },
110
+ {
111
+ "reward_call": 3,
112
+ "index": 1,
113
+ "score": -0.6649999999999999,
114
+ "task_id": "bike_accessory_mount_120n",
115
+ "code_chars": 6274,
116
+ "cadforge_build": 0.0,
117
+ "cadforge_total": -1.0,
118
+ "has_fixture": true,
119
+ "has_cadquery_import": true,
120
+ "error_kind": "SyntaxError"
121
+ },
122
+ {
123
+ "reward_call": 3,
124
+ "index": 2,
125
+ "score": 0.7981208947102704,
126
+ "task_id": "laptop_riser_vented",
127
+ "code_chars": 3990,
128
+ "cadforge_build": 1.0,
129
+ "cadforge_total": 0.7624951702473769,
130
+ "has_fixture": true,
131
+ "has_cadquery_import": true,
132
+ "error_kind": "build_ok"
133
+ },
134
+ {
135
+ "reward_call": 3,
136
+ "index": 3,
137
+ "score": -0.84,
138
+ "task_id": "laptop_riser_vented",
139
+ "code_chars": 5472,
140
+ "cadforge_build": 0.0,
141
+ "cadforge_total": -1.0,
142
+ "has_fixture": false,
143
+ "has_cadquery_import": true,
144
+ "error_kind": "SyntaxError"
145
+ },
146
+ {
147
+ "reward_call": 4,
148
+ "index": 0,
149
+ "score": -0.84,
150
+ "task_id": "bench_vise_simplified",
151
+ "code_chars": 5334,
152
+ "cadforge_build": 0.0,
153
+ "cadforge_total": -1.0,
154
+ "has_fixture": false,
155
+ "has_cadquery_import": true,
156
+ "error_kind": "SyntaxError"
157
+ },
158
+ {
159
+ "reward_call": 4,
160
+ "index": 1,
161
+ "score": -0.84,
162
+ "task_id": "bench_vise_simplified",
163
+ "code_chars": 5572,
164
+ "cadforge_build": 0.0,
165
+ "cadforge_total": -1.0,
166
+ "has_fixture": false,
167
+ "has_cadquery_import": true,
168
+ "error_kind": "SyntaxError"
169
+ },
170
+ {
171
+ "reward_call": 4,
172
+ "index": 2,
173
+ "score": 0.6571400489319114,
174
+ "task_id": "desk_lamp_arm_joint",
175
+ "code_chars": 5885,
176
+ "cadforge_build": 1.0,
177
+ "cadforge_total": 0.5966353516846016,
178
+ "has_fixture": true,
179
+ "has_cadquery_import": true,
180
+ "error_kind": "build_ok"
181
+ },
182
+ {
183
+ "reward_call": 4,
184
+ "index": 3,
185
+ "score": -0.84,
186
+ "task_id": "desk_lamp_arm_joint",
187
+ "code_chars": 5805,
188
+ "cadforge_build": 0.0,
189
+ "cadforge_total": -1.0,
190
+ "has_fixture": false,
191
+ "has_cadquery_import": true,
192
+ "error_kind": "SyntaxError"
193
+ },
194
+ {
195
+ "reward_call": 5,
196
+ "index": 0,
197
+ "score": -0.74,
198
+ "task_id": "ergonomic_curvy_chair_1000n",
199
+ "code_chars": 5169,
200
+ "cadforge_build": 0.0,
201
+ "cadforge_total": -1.0,
202
+ "has_fixture": false,
203
+ "has_cadquery_import": true,
204
+ "error_kind": "TypeError"
205
+ },
206
+ {
207
+ "reward_call": 5,
208
+ "index": 1,
209
+ "score": -0.84,
210
+ "task_id": "ergonomic_curvy_chair_1000n",
211
+ "code_chars": 5408,
212
+ "cadforge_build": 0.0,
213
+ "cadforge_total": -1.0,
214
+ "has_fixture": false,
215
+ "has_cadquery_import": true,
216
+ "error_kind": "SyntaxError"
217
+ },
218
+ {
219
+ "reward_call": 5,
220
+ "index": 2,
221
+ "score": -0.645,
222
+ "task_id": "lightweight_truss_support_250n",
223
+ "code_chars": 4484,
224
+ "cadforge_build": 0.0,
225
+ "cadforge_total": -1.0,
226
+ "has_fixture": true,
227
+ "has_cadquery_import": true,
228
+ "error_kind": "AttributeError"
229
+ },
230
+ {
231
+ "reward_call": 5,
232
+ "index": 3,
233
+ "score": -0.61,
234
+ "task_id": "lightweight_truss_support_250n",
235
+ "code_chars": 3554,
236
+ "cadforge_build": 0.0,
237
+ "cadforge_total": -1.0,
238
+ "has_fixture": true,
239
+ "has_cadquery_import": true,
240
+ "error_kind": "ValueError"
241
+ },
242
+ {
243
+ "reward_call": 6,
244
+ "index": 0,
245
+ "score": -0.84,
246
+ "task_id": "gearbox_housing_split",
247
+ "code_chars": 5514,
248
+ "cadforge_build": 0.0,
249
+ "cadforge_total": -1.0,
250
+ "has_fixture": false,
251
+ "has_cadquery_import": true,
252
+ "error_kind": "SyntaxError"
253
+ },
254
+ {
255
+ "reward_call": 6,
256
+ "index": 1,
257
+ "score": -0.865,
258
+ "task_id": "gearbox_housing_split",
259
+ "code_chars": 5454,
260
+ "cadforge_build": 0.0,
261
+ "cadforge_total": -1.0,
262
+ "has_fixture": false,
263
+ "has_cadquery_import": true,
264
+ "error_kind": "AttributeError"
265
+ },
266
+ {
267
+ "reward_call": 6,
268
+ "index": 2,
269
+ "score": -0.6649999999999999,
270
+ "task_id": "cable_clip_snap",
271
+ "code_chars": 4792,
272
+ "cadforge_build": 0.0,
273
+ "cadforge_total": -1.0,
274
+ "has_fixture": true,
275
+ "has_cadquery_import": true,
276
+ "error_kind": "SyntaxError"
277
+ },
278
+ {
279
+ "reward_call": 6,
280
+ "index": 3,
281
+ "score": -0.84,
282
+ "task_id": "cable_clip_snap",
283
+ "code_chars": 5814,
284
+ "cadforge_build": 0.0,
285
+ "cadforge_total": -1.0,
286
+ "has_fixture": false,
287
+ "has_cadquery_import": true,
288
+ "error_kind": "SyntaxError"
289
+ },
290
+ {
291
+ "reward_call": 7,
292
+ "index": 0,
293
+ "score": -0.84,
294
+ "task_id": "folding_step_stool",
295
+ "code_chars": 6285,
296
+ "cadforge_build": 0.0,
297
+ "cadforge_total": -1.0,
298
+ "has_fixture": false,
299
+ "has_cadquery_import": true,
300
+ "error_kind": "SyntaxError"
301
+ },
302
+ {
303
+ "reward_call": 7,
304
+ "index": 1,
305
+ "score": -0.84,
306
+ "task_id": "folding_step_stool",
307
+ "code_chars": 5364,
308
+ "cadforge_build": 0.0,
309
+ "cadforge_total": -1.0,
310
+ "has_fixture": false,
311
+ "has_cadquery_import": true,
312
+ "error_kind": "SyntaxError"
313
+ },
314
+ {
315
+ "reward_call": 7,
316
+ "index": 2,
317
+ "score": 0.6945054503947041,
318
+ "task_id": "shelf_bracket_triangular_200n",
319
+ "code_chars": 3450,
320
+ "cadforge_build": 1.0,
321
+ "cadforge_total": 0.6405946475231813,
322
+ "has_fixture": true,
323
+ "has_cadquery_import": true,
324
+ "error_kind": "build_ok"
325
+ },
326
+ {
327
+ "reward_call": 7,
328
+ "index": 3,
329
+ "score": -0.565,
330
+ "task_id": "shelf_bracket_triangular_200n",
331
+ "code_chars": 3749,
332
+ "cadforge_build": 0.0,
333
+ "cadforge_total": -1.0,
334
+ "has_fixture": true,
335
+ "has_cadquery_import": true,
336
+ "error_kind": "TypeError"
337
+ },
338
+ {
339
+ "reward_call": 8,
340
+ "index": 0,
341
+ "score": -0.565,
342
+ "task_id": "phone_stand_adjustable_hinge",
343
+ "code_chars": 4750,
344
+ "cadforge_build": 0.0,
345
+ "cadforge_total": -1.0,
346
+ "has_fixture": true,
347
+ "has_cadquery_import": true,
348
+ "error_kind": "ValueError"
349
+ },
350
+ {
351
+ "reward_call": 8,
352
+ "index": 1,
353
+ "score": 0.8989119400703898,
354
+ "task_id": "phone_stand_adjustable_hinge",
355
+ "code_chars": 4176,
356
+ "cadforge_build": 1.0,
357
+ "cadforge_total": 0.8810728706710468,
358
+ "has_fixture": true,
359
+ "has_cadquery_import": true,
360
+ "error_kind": "build_ok"
361
+ },
362
+ {
363
+ "reward_call": 8,
364
+ "index": 2,
365
+ "score": -0.565,
366
+ "task_id": "micro_drone_frame",
367
+ "code_chars": 5268,
368
+ "cadforge_build": 0.0,
369
+ "cadforge_total": -1.0,
370
+ "has_fixture": true,
371
+ "has_cadquery_import": true,
372
+ "error_kind": "TypeError"
373
+ },
374
+ {
375
+ "reward_call": 8,
376
+ "index": 3,
377
+ "score": -0.84,
378
+ "task_id": "micro_drone_frame",
379
+ "code_chars": 5917,
380
+ "cadforge_build": 0.0,
381
+ "cadforge_total": -1.0,
382
+ "has_fixture": false,
383
+ "has_cadquery_import": true,
384
+ "error_kind": "SyntaxError"
385
+ },
386
+ {
387
+ "reward_call": 9,
388
+ "index": 0,
389
+ "score": -0.6649999999999999,
390
+ "task_id": "wall_j_hook_120n",
391
+ "code_chars": 3619,
392
+ "cadforge_build": 0.0,
393
+ "cadforge_total": -1.0,
394
+ "has_fixture": true,
395
+ "has_cadquery_import": true,
396
+ "error_kind": "SyntaxError"
397
+ },
398
+ {
399
+ "reward_call": 9,
400
+ "index": 1,
401
+ "score": -0.565,
402
+ "task_id": "wall_j_hook_120n",
403
+ "code_chars": 4369,
404
+ "cadforge_build": 0.0,
405
+ "cadforge_total": -1.0,
406
+ "has_fixture": true,
407
+ "has_cadquery_import": true,
408
+ "error_kind": "TypeError"
409
+ },
410
+ {
411
+ "reward_call": 9,
412
+ "index": 2,
413
+ "score": -0.565,
414
+ "task_id": "drawer_handle_two_bolt",
415
+ "code_chars": 3920,
416
+ "cadforge_build": 0.0,
417
+ "cadforge_total": -1.0,
418
+ "has_fixture": true,
419
+ "has_cadquery_import": true,
420
+ "error_kind": "TypeError"
421
+ },
422
+ {
423
+ "reward_call": 9,
424
+ "index": 3,
425
+ "score": 0.7152374233607848,
426
+ "task_id": "drawer_handle_two_bolt",
427
+ "code_chars": 2282,
428
+ "cadforge_build": 1.0,
429
+ "cadforge_total": 0.6649852039538644,
430
+ "has_fixture": true,
431
+ "has_cadquery_import": true,
432
+ "error_kind": "build_ok"
433
+ },
434
+ {
435
+ "reward_call": 10,
436
+ "index": 0,
437
+ "score": -0.84,
438
+ "task_id": "shaft_torque_clamp_120nm",
439
+ "code_chars": 5215,
440
+ "cadforge_build": 0.0,
441
+ "cadforge_total": -1.0,
442
+ "has_fixture": true,
443
+ "has_cadquery_import": true,
444
+ "error_kind": "SyntaxError"
445
+ },
446
+ {
447
+ "reward_call": 10,
448
+ "index": 1,
449
+ "score": -0.615,
450
+ "task_id": "shaft_torque_clamp_120nm",
451
+ "code_chars": 5209,
452
+ "cadforge_build": 0.0,
453
+ "cadforge_total": -1.0,
454
+ "has_fixture": true,
455
+ "has_cadquery_import": true,
456
+ "error_kind": "NameError"
457
+ },
458
+ {
459
+ "reward_call": 10,
460
+ "index": 2,
461
+ "score": -0.595,
462
+ "task_id": "hinge_leaf_barrel",
463
+ "code_chars": 4837,
464
+ "cadforge_build": 0.0,
465
+ "cadforge_total": -1.0,
466
+ "has_fixture": true,
467
+ "has_cadquery_import": true,
468
+ "error_kind": "BRep_API"
469
+ },
470
+ {
471
+ "reward_call": 10,
472
+ "index": 3,
473
+ "score": -0.595,
474
+ "task_id": "hinge_leaf_barrel",
475
+ "code_chars": 6057,
476
+ "cadforge_build": 0.0,
477
+ "cadforge_total": -1.0,
478
+ "has_fixture": true,
479
+ "has_cadquery_import": true,
480
+ "error_kind": "ValueError"
481
+ },
482
+ {
483
+ "reward_call": 11,
484
+ "index": 0,
485
+ "score": -0.79,
486
+ "task_id": "ergonomic_curvy_chair_1000n",
487
+ "code_chars": 4883,
488
+ "cadforge_build": 0.0,
489
+ "cadforge_total": -1.0,
490
+ "has_fixture": false,
491
+ "has_cadquery_import": true,
492
+ "error_kind": "AttributeError"
493
+ },
494
+ {
495
+ "reward_call": 11,
496
+ "index": 1,
497
+ "score": -0.84,
498
+ "task_id": "ergonomic_curvy_chair_1000n",
499
+ "code_chars": 5198,
500
+ "cadforge_build": 0.0,
501
+ "cadforge_total": -1.0,
502
+ "has_fixture": false,
503
+ "has_cadquery_import": true,
504
+ "error_kind": "SyntaxError"
505
+ },
506
+ {
507
+ "reward_call": 11,
508
+ "index": 2,
509
+ "score": 0.7985873183029124,
510
+ "task_id": "pipe_clamp_saddle",
511
+ "code_chars": 4307,
512
+ "cadforge_build": 1.0,
513
+ "cadforge_total": 0.7630439038857793,
514
+ "has_fixture": true,
515
+ "has_cadquery_import": true,
516
+ "error_kind": "build_ok"
517
+ },
518
+ {
519
+ "reward_call": 11,
520
+ "index": 3,
521
+ "score": -0.61,
522
+ "task_id": "pipe_clamp_saddle",
523
+ "code_chars": 4098,
524
+ "cadforge_build": 0.0,
525
+ "cadforge_total": -1.0,
526
+ "has_fixture": true,
527
+ "has_cadquery_import": true,
528
+ "error_kind": "unknown"
529
+ },
530
+ {
531
+ "reward_call": 12,
532
+ "index": 0,
533
+ "score": -0.84,
534
+ "task_id": "laptop_riser_vented",
535
+ "code_chars": 5792,
536
+ "cadforge_build": 0.0,
537
+ "cadforge_total": -1.0,
538
+ "has_fixture": false,
539
+ "has_cadquery_import": true,
540
+ "error_kind": "SyntaxError"
541
+ },
542
+ {
543
+ "reward_call": 12,
544
+ "index": 1,
545
+ "score": 0.8229701323180416,
546
+ "task_id": "laptop_riser_vented",
547
+ "code_chars": 4243,
548
+ "cadforge_build": 1.0,
549
+ "cadforge_total": 0.7917295674329902,
550
+ "has_fixture": true,
551
+ "has_cadquery_import": true,
552
+ "error_kind": "build_ok"
553
+ },
554
+ {
555
+ "reward_call": 12,
556
+ "index": 2,
557
+ "score": -0.84,
558
+ "task_id": "gearbox_housing_split",
559
+ "code_chars": 5805,
560
+ "cadforge_build": 0.0,
561
+ "cadforge_total": -1.0,
562
+ "has_fixture": false,
563
+ "has_cadquery_import": true,
564
+ "error_kind": "SyntaxError"
565
+ },
566
+ {
567
+ "reward_call": 12,
568
+ "index": 3,
569
+ "score": -0.84,
570
+ "task_id": "gearbox_housing_split",
571
+ "code_chars": 5905,
572
+ "cadforge_build": 0.0,
573
+ "cadforge_total": -1.0,
574
+ "has_fixture": false,
575
+ "has_cadquery_import": true,
576
+ "error_kind": "SyntaxError"
577
+ },
578
+ {
579
+ "reward_call": 13,
580
+ "index": 0,
581
+ "score": -0.565,
582
+ "task_id": "lightweight_truss_support_250n",
583
+ "code_chars": 3976,
584
+ "cadforge_build": 0.0,
585
+ "cadforge_total": -1.0,
586
+ "has_fixture": true,
587
+ "has_cadquery_import": true,
588
+ "error_kind": "TypeError"
589
+ },
590
+ {
591
+ "reward_call": 13,
592
+ "index": 1,
593
+ "score": -0.8699999999999999,
594
+ "task_id": "lightweight_truss_support_250n",
595
+ "code_chars": 5353,
596
+ "cadforge_build": 0.0,
597
+ "cadforge_total": -1.0,
598
+ "has_fixture": false,
599
+ "has_cadquery_import": true,
600
+ "error_kind": "SyntaxError"
601
+ },
602
+ {
603
+ "reward_call": 13,
604
+ "index": 2,
605
+ "score": 0.5920885416666667,
606
+ "task_id": "wall_j_hook_120n",
607
+ "code_chars": 2571,
608
+ "cadforge_build": 1.0,
609
+ "cadforge_total": 0.5201041666666667,
610
+ "has_fixture": true,
611
+ "has_cadquery_import": true,
612
+ "error_kind": "build_ok"
613
+ },
614
+ {
615
+ "reward_call": 13,
616
+ "index": 3,
617
+ "score": -0.615,
618
+ "task_id": "wall_j_hook_120n",
619
+ "code_chars": 4179,
620
+ "cadforge_build": 0.0,
621
+ "cadforge_total": -1.0,
622
+ "has_fixture": true,
623
+ "has_cadquery_import": true,
624
+ "error_kind": "AttributeError"
625
+ },
626
+ {
627
+ "reward_call": 14,
628
+ "index": 0,
629
+ "score": -0.84,
630
+ "task_id": "desk_lamp_arm_joint",
631
+ "code_chars": 5644,
632
+ "cadforge_build": 0.0,
633
+ "cadforge_total": -1.0,
634
+ "has_fixture": false,
635
+ "has_cadquery_import": true,
636
+ "error_kind": "SyntaxError"
637
+ },
638
+ {
639
+ "reward_call": 14,
640
+ "index": 1,
641
+ "score": -0.84,
642
+ "task_id": "desk_lamp_arm_joint",
643
+ "code_chars": 5694,
644
+ "cadforge_build": 0.0,
645
+ "cadforge_total": -1.0,
646
+ "has_fixture": false,
647
+ "has_cadquery_import": true,
648
+ "error_kind": "SyntaxError"
649
+ },
650
+ {
651
+ "reward_call": 14,
652
+ "index": 2,
653
+ "score": 0.7292988990436297,
654
+ "task_id": "robot_gripper_two_finger",
655
+ "code_chars": 4459,
656
+ "cadforge_build": 1.0,
657
+ "cadforge_total": 0.6815281165219172,
658
+ "has_fixture": true,
659
+ "has_cadquery_import": true,
660
+ "error_kind": "build_ok"
661
+ },
662
+ {
663
+ "reward_call": 14,
664
+ "index": 3,
665
+ "score": -0.6649999999999999,
666
+ "task_id": "robot_gripper_two_finger",
667
+ "code_chars": 5451,
668
+ "cadforge_build": 0.0,
669
+ "cadforge_total": -1.0,
670
+ "has_fixture": true,
671
+ "has_cadquery_import": true,
672
+ "error_kind": "SyntaxError"
673
+ },
674
+ {
675
+ "reward_call": 15,
676
+ "index": 0,
677
+ "score": 0.7945833333333334,
678
+ "task_id": "shelf_bracket_triangular_200n",
679
+ "code_chars": 4031,
680
+ "cadforge_build": 1.0,
681
+ "cadforge_total": 0.7583333333333334,
682
+ "has_fixture": true,
683
+ "has_cadquery_import": true,
684
+ "error_kind": "build_ok"
685
+ },
686
+ {
687
+ "reward_call": 15,
688
+ "index": 1,
689
+ "score": 0.8232339705474382,
690
+ "task_id": "shelf_bracket_triangular_200n",
691
+ "code_chars": 3850,
692
+ "cadforge_build": 1.0,
693
+ "cadforge_total": 0.7920399653499273,
694
+ "has_fixture": true,
695
+ "has_cadquery_import": true,
696
+ "error_kind": "build_ok"
697
+ },
698
+ {
699
+ "reward_call": 15,
700
+ "index": 2,
701
+ "score": -0.84,
702
+ "task_id": "micro_drone_frame",
703
+ "code_chars": 6013,
704
+ "cadforge_build": 0.0,
705
+ "cadforge_total": -1.0,
706
+ "has_fixture": false,
707
+ "has_cadquery_import": true,
708
+ "error_kind": "SyntaxError"
709
+ },
710
+ {
711
+ "reward_call": 15,
712
+ "index": 3,
713
+ "score": -0.84,
714
+ "task_id": "micro_drone_frame",
715
+ "code_chars": 5922,
716
+ "cadforge_build": 0.0,
717
+ "cadforge_total": -1.0,
718
+ "has_fixture": false,
719
+ "has_cadquery_import": true,
720
+ "error_kind": "SyntaxError"
721
+ },
722
+ {
723
+ "reward_call": 16,
724
+ "index": 0,
725
+ "score": -0.595,
726
+ "task_id": "drawer_handle_two_bolt",
727
+ "code_chars": 4798,
728
+ "cadforge_build": 0.0,
729
+ "cadforge_total": -1.0,
730
+ "has_fixture": true,
731
+ "has_cadquery_import": true,
732
+ "error_kind": "unknown"
733
+ },
734
+ {
735
+ "reward_call": 16,
736
+ "index": 1,
737
+ "score": 0.7621206394200344,
738
+ "task_id": "drawer_handle_two_bolt",
739
+ "code_chars": 2362,
740
+ "cadforge_build": 1.0,
741
+ "cadforge_total": 0.7201419287294522,
742
+ "has_fixture": true,
743
+ "has_cadquery_import": true,
744
+ "error_kind": "build_ok"
745
+ },
746
+ {
747
+ "reward_call": 16,
748
+ "index": 2,
749
+ "score": 0.6708019450756568,
750
+ "task_id": "bike_accessory_mount_120n",
751
+ "code_chars": 3691,
752
+ "cadforge_build": 1.0,
753
+ "cadforge_total": 0.6127081706772433,
754
+ "has_fixture": true,
755
+ "has_cadquery_import": true,
756
+ "error_kind": "build_ok"
757
+ },
758
+ {
759
+ "reward_call": 16,
760
+ "index": 3,
761
+ "score": 0.748018455196793,
762
+ "task_id": "bike_accessory_mount_120n",
763
+ "code_chars": 4511,
764
+ "cadforge_build": 1.0,
765
+ "cadforge_total": 0.7035511237609329,
766
+ "has_fixture": true,
767
+ "has_cadquery_import": true,
768
+ "error_kind": "build_ok"
769
+ },
770
+ {
771
+ "reward_call": 17,
772
+ "index": 0,
773
+ "score": -0.595,
774
+ "task_id": "hinge_leaf_barrel",
775
+ "code_chars": 5141,
776
+ "cadforge_build": 0.0,
777
+ "cadforge_total": -1.0,
778
+ "has_fixture": true,
779
+ "has_cadquery_import": true,
780
+ "error_kind": "ValueError"
781
+ },
782
+ {
783
+ "reward_call": 17,
784
+ "index": 1,
785
+ "score": 0.6935120777921006,
786
+ "task_id": "hinge_leaf_barrel",
787
+ "code_chars": 5678,
788
+ "cadforge_build": 1.0,
789
+ "cadforge_total": 0.6394259738730594,
790
+ "has_fixture": true,
791
+ "has_cadquery_import": true,
792
+ "error_kind": "build_ok"
793
+ },
794
+ {
795
+ "reward_call": 17,
796
+ "index": 2,
797
+ "score": -0.84,
798
+ "task_id": "phone_stand_adjustable_hinge",
799
+ "code_chars": 5903,
800
+ "cadforge_build": 0.0,
801
+ "cadforge_total": -1.0,
802
+ "has_fixture": false,
803
+ "has_cadquery_import": true,
804
+ "error_kind": "SyntaxError"
805
+ },
806
+ {
807
+ "reward_call": 17,
808
+ "index": 3,
809
+ "score": 0.8979407835056935,
810
+ "task_id": "phone_stand_adjustable_hinge",
811
+ "code_chars": 3846,
812
+ "cadforge_build": 1.0,
813
+ "cadforge_total": 0.87993033353611,
814
+ "has_fixture": true,
815
+ "has_cadquery_import": true,
816
+ "error_kind": "build_ok"
817
+ },
818
+ {
819
+ "reward_call": 18,
820
+ "index": 0,
821
+ "score": -0.84,
822
+ "task_id": "folding_step_stool",
823
+ "code_chars": 5335,
824
+ "cadforge_build": 0.0,
825
+ "cadforge_total": -1.0,
826
+ "has_fixture": false,
827
+ "has_cadquery_import": true,
828
+ "error_kind": "SyntaxError"
829
+ },
830
+ {
831
+ "reward_call": 18,
832
+ "index": 1,
833
+ "score": -0.84,
834
+ "task_id": "folding_step_stool",
835
+ "code_chars": 5235,
836
+ "cadforge_build": 0.0,
837
+ "cadforge_total": -1.0,
838
+ "has_fixture": false,
839
+ "has_cadquery_import": true,
840
+ "error_kind": "SyntaxError"
841
+ },
842
+ {
843
+ "reward_call": 18,
844
+ "index": 2,
845
+ "score": -0.615,
846
+ "task_id": "table_six_leg_500n",
847
+ "code_chars": 3404,
848
+ "cadforge_build": 0.0,
849
+ "cadforge_total": -1.0,
850
+ "has_fixture": true,
851
+ "has_cadquery_import": true,
852
+ "error_kind": "AttributeError"
853
+ },
854
+ {
855
+ "reward_call": 18,
856
+ "index": 3,
857
+ "score": 0.8017035468139491,
858
+ "task_id": "table_six_leg_500n",
859
+ "code_chars": 5282,
860
+ "cadforge_build": 1.0,
861
+ "cadforge_total": 0.7667100550752342,
862
+ "has_fixture": true,
863
+ "has_cadquery_import": true,
864
+ "error_kind": "build_ok"
865
+ },
866
+ {
867
+ "reward_call": 19,
868
+ "index": 0,
869
+ "score": 0.5738930435982795,
870
+ "task_id": "shaft_torque_clamp_120nm",
871
+ "code_chars": 3560,
872
+ "cadforge_build": 1.0,
873
+ "cadforge_total": 0.4986976983509171,
874
+ "has_fixture": true,
875
+ "has_cadquery_import": true,
876
+ "error_kind": "build_ok"
877
+ },
878
+ {
879
+ "reward_call": 19,
880
+ "index": 1,
881
+ "score": -0.6649999999999999,
882
+ "task_id": "shaft_torque_clamp_120nm",
883
+ "code_chars": 5936,
884
+ "cadforge_build": 0.0,
885
+ "cadforge_total": -1.0,
886
+ "has_fixture": true,
887
+ "has_cadquery_import": true,
888
+ "error_kind": "SyntaxError"
889
+ },
890
+ {
891
+ "reward_call": 19,
892
+ "index": 2,
893
+ "score": -0.565,
894
+ "task_id": "bottle_cage_bike",
895
+ "code_chars": 5253,
896
+ "cadforge_build": 0.0,
897
+ "cadforge_total": -1.0,
898
+ "has_fixture": true,
899
+ "has_cadquery_import": true,
900
+ "error_kind": "TypeError"
901
+ },
902
+ {
903
+ "reward_call": 19,
904
+ "index": 3,
905
+ "score": -0.565,
906
+ "task_id": "bottle_cage_bike",
907
+ "code_chars": 5862,
908
+ "cadforge_build": 0.0,
909
+ "cadforge_total": -1.0,
910
+ "has_fixture": true,
911
+ "has_cadquery_import": true,
912
+ "error_kind": "unknown"
913
+ },
914
+ {
915
+ "reward_call": 20,
916
+ "index": 0,
917
+ "score": -0.71,
918
+ "task_id": "bench_vise_simplified",
919
+ "code_chars": 4991,
920
+ "cadforge_build": 0.0,
921
+ "cadforge_total": -1.0,
922
+ "has_fixture": true,
923
+ "has_cadquery_import": true,
924
+ "error_kind": "SyntaxError"
925
+ },
926
+ {
927
+ "reward_call": 20,
928
+ "index": 1,
929
+ "score": -0.84,
930
+ "task_id": "bench_vise_simplified",
931
+ "code_chars": 5288,
932
+ "cadforge_build": 0.0,
933
+ "cadforge_total": -1.0,
934
+ "has_fixture": false,
935
+ "has_cadquery_import": true,
936
+ "error_kind": "SyntaxError"
937
+ },
938
+ {
939
+ "reward_call": 20,
940
+ "index": 2,
941
+ "score": 0.8406452528509025,
942
+ "task_id": "cable_clip_snap",
943
+ "code_chars": 4185,
944
+ "cadforge_build": 1.0,
945
+ "cadforge_total": 0.8125238268834147,
946
+ "has_fixture": true,
947
+ "has_cadquery_import": true,
948
+ "error_kind": "build_ok"
949
+ },
950
+ {
951
+ "reward_call": 20,
952
+ "index": 3,
953
+ "score": -0.565,
954
+ "task_id": "cable_clip_snap",
955
+ "code_chars": 5716,
956
+ "cadforge_build": 0.0,
957
+ "cadforge_total": -1.0,
958
+ "has_fixture": true,
959
+ "has_cadquery_import": true,
960
+ "error_kind": "BRep_API"
961
+ },
962
+ {
963
+ "reward_call": 21,
964
+ "index": 0,
965
+ "score": -0.565,
966
+ "task_id": "micro_drone_frame",
967
+ "code_chars": 5690,
968
+ "cadforge_build": 0.0,
969
+ "cadforge_total": -1.0,
970
+ "has_fixture": true,
971
+ "has_cadquery_import": true,
972
+ "error_kind": "ValueError"
973
+ },
974
+ {
975
+ "reward_call": 21,
976
+ "index": 1,
977
+ "score": -0.615,
978
+ "task_id": "micro_drone_frame",
979
+ "code_chars": 5600,
980
+ "cadforge_build": 0.0,
981
+ "cadforge_total": -1.0,
982
+ "has_fixture": true,
983
+ "has_cadquery_import": true,
984
+ "error_kind": "NameError"
985
+ },
986
+ {
987
+ "reward_call": 21,
988
+ "index": 2,
989
+ "score": 0.6909928797128871,
990
+ "task_id": "bike_accessory_mount_120n",
991
+ "code_chars": 3893,
992
+ "cadforge_build": 1.0,
993
+ "cadforge_total": 0.636462211426926,
994
+ "has_fixture": true,
995
+ "has_cadquery_import": true,
996
+ "error_kind": "build_ok"
997
+ },
998
+ {
999
+ "reward_call": 21,
1000
+ "index": 3,
1001
+ "score": -0.615,
1002
+ "task_id": "bike_accessory_mount_120n",
1003
+ "code_chars": 4698,
1004
+ "cadforge_build": 0.0,
1005
+ "cadforge_total": -1.0,
1006
+ "has_fixture": true,
1007
+ "has_cadquery_import": true,
1008
+ "error_kind": "AttributeError"
1009
+ },
1010
+ {
1011
+ "reward_call": 22,
1012
+ "index": 0,
1013
+ "score": -0.84,
1014
+ "task_id": "desk_lamp_arm_joint",
1015
+ "code_chars": 5290,
1016
+ "cadforge_build": 0.0,
1017
+ "cadforge_total": -1.0,
1018
+ "has_fixture": false,
1019
+ "has_cadquery_import": true,
1020
+ "error_kind": "SyntaxError"
1021
+ },
1022
+ {
1023
+ "reward_call": 22,
1024
+ "index": 1,
1025
+ "score": -0.84,
1026
+ "task_id": "desk_lamp_arm_joint",
1027
+ "code_chars": 6092,
1028
+ "cadforge_build": 0.0,
1029
+ "cadforge_total": -1.0,
1030
+ "has_fixture": false,
1031
+ "has_cadquery_import": true,
1032
+ "error_kind": "SyntaxError"
1033
+ },
1034
+ {
1035
+ "reward_call": 22,
1036
+ "index": 2,
1037
+ "score": 0.7858514119963492,
1038
+ "task_id": "robot_gripper_two_finger",
1039
+ "code_chars": 3520,
1040
+ "cadforge_build": 1.0,
1041
+ "cadforge_total": 0.7480604847015874,
1042
+ "has_fixture": true,
1043
+ "has_cadquery_import": true,
1044
+ "error_kind": "build_ok"
1045
+ },
1046
+ {
1047
+ "reward_call": 22,
1048
+ "index": 3,
1049
+ "score": -0.84,
1050
+ "task_id": "robot_gripper_two_finger",
1051
+ "code_chars": 5443,
1052
+ "cadforge_build": 0.0,
1053
+ "cadforge_total": -1.0,
1054
+ "has_fixture": false,
1055
+ "has_cadquery_import": true,
1056
+ "error_kind": "SyntaxError"
1057
+ },
1058
+ {
1059
+ "reward_call": 23,
1060
+ "index": 0,
1061
+ "score": -0.565,
1062
+ "task_id": "phone_stand_adjustable_hinge",
1063
+ "code_chars": 5380,
1064
+ "cadforge_build": 0.0,
1065
+ "cadforge_total": -1.0,
1066
+ "has_fixture": true,
1067
+ "has_cadquery_import": true,
1068
+ "error_kind": "TypeError"
1069
+ },
1070
+ {
1071
+ "reward_call": 23,
1072
+ "index": 1,
1073
+ "score": 0.8120360233537414,
1074
+ "task_id": "phone_stand_adjustable_hinge",
1075
+ "code_chars": 5196,
1076
+ "cadforge_build": 1.0,
1077
+ "cadforge_total": 0.778865909827931,
1078
+ "has_fixture": true,
1079
+ "has_cadquery_import": true,
1080
+ "error_kind": "build_ok"
1081
+ },
1082
+ {
1083
+ "reward_call": 23,
1084
+ "index": 2,
1085
+ "score": -0.6649999999999999,
1086
+ "task_id": "pipe_clamp_saddle",
1087
+ "code_chars": 3213,
1088
+ "cadforge_build": 0.0,
1089
+ "cadforge_total": -1.0,
1090
+ "has_fixture": true,
1091
+ "has_cadquery_import": true,
1092
+ "error_kind": "SyntaxError"
1093
+ },
1094
+ {
1095
+ "reward_call": 23,
1096
+ "index": 3,
1097
+ "score": -0.84,
1098
+ "task_id": "pipe_clamp_saddle",
1099
+ "code_chars": 5624,
1100
+ "cadforge_build": 0.0,
1101
+ "cadforge_total": -1.0,
1102
+ "has_fixture": false,
1103
+ "has_cadquery_import": true,
1104
+ "error_kind": "SyntaxError"
1105
+ },
1106
+ {
1107
+ "reward_call": 24,
1108
+ "index": 0,
1109
+ "score": 0.7216417724094514,
1110
+ "task_id": "wall_j_hook_120n",
1111
+ "code_chars": 2812,
1112
+ "cadforge_build": 1.0,
1113
+ "cadforge_total": 0.6725197322464134,
1114
+ "has_fixture": true,
1115
+ "has_cadquery_import": true,
1116
+ "error_kind": "build_ok"
1117
+ },
1118
+ {
1119
+ "reward_call": 24,
1120
+ "index": 1,
1121
+ "score": -0.615,
1122
+ "task_id": "wall_j_hook_120n",
1123
+ "code_chars": 4168,
1124
+ "cadforge_build": 0.0,
1125
+ "cadforge_total": -1.0,
1126
+ "has_fixture": true,
1127
+ "has_cadquery_import": true,
1128
+ "error_kind": "NameError"
1129
+ },
1130
+ {
1131
+ "reward_call": 24,
1132
+ "index": 2,
1133
+ "score": -0.6649999999999999,
1134
+ "task_id": "gearbox_housing_split",
1135
+ "code_chars": 5075,
1136
+ "cadforge_build": 0.0,
1137
+ "cadforge_total": -1.0,
1138
+ "has_fixture": true,
1139
+ "has_cadquery_import": true,
1140
+ "error_kind": "SyntaxError"
1141
+ },
1142
+ {
1143
+ "reward_call": 24,
1144
+ "index": 3,
1145
+ "score": -0.84,
1146
+ "task_id": "gearbox_housing_split",
1147
+ "code_chars": 5260,
1148
+ "cadforge_build": 0.0,
1149
+ "cadforge_total": -1.0,
1150
+ "has_fixture": false,
1151
+ "has_cadquery_import": true,
1152
+ "error_kind": "SyntaxError"
1153
+ },
1154
+ {
1155
+ "reward_call": 25,
1156
+ "index": 0,
1157
+ "score": -0.84,
1158
+ "task_id": "ergonomic_curvy_chair_1000n",
1159
+ "code_chars": 5532,
1160
+ "cadforge_build": 0.0,
1161
+ "cadforge_total": -1.0,
1162
+ "has_fixture": false,
1163
+ "has_cadquery_import": true,
1164
+ "error_kind": "SyntaxError"
1165
+ },
1166
+ {
1167
+ "reward_call": 25,
1168
+ "index": 1,
1169
+ "score": -0.84,
1170
+ "task_id": "ergonomic_curvy_chair_1000n",
1171
+ "code_chars": 4912,
1172
+ "cadforge_build": 0.0,
1173
+ "cadforge_total": -1.0,
1174
+ "has_fixture": false,
1175
+ "has_cadquery_import": true,
1176
+ "error_kind": "SyntaxError"
1177
+ },
1178
+ {
1179
+ "reward_call": 25,
1180
+ "index": 2,
1181
+ "score": 0.8267267120260791,
1182
+ "task_id": "table_six_leg_500n",
1183
+ "code_chars": 3063,
1184
+ "cadforge_build": 1.0,
1185
+ "cadforge_total": 0.7961490729718578,
1186
+ "has_fixture": true,
1187
+ "has_cadquery_import": true,
1188
+ "error_kind": "build_ok"
1189
+ },
1190
+ {
1191
+ "reward_call": 25,
1192
+ "index": 3,
1193
+ "score": 0.7771561208338964,
1194
+ "task_id": "table_six_leg_500n",
1195
+ "code_chars": 5640,
1196
+ "cadforge_build": 1.0,
1197
+ "cadforge_total": 0.7378307303928193,
1198
+ "has_fixture": true,
1199
+ "has_cadquery_import": true,
1200
+ "error_kind": "build_ok"
1201
+ },
1202
+ {
1203
+ "reward_call": 26,
1204
+ "index": 0,
1205
+ "score": -0.565,
1206
+ "task_id": "bottle_cage_bike",
1207
+ "code_chars": 5659,
1208
+ "cadforge_build": 0.0,
1209
+ "cadforge_total": -1.0,
1210
+ "has_fixture": true,
1211
+ "has_cadquery_import": true,
1212
+ "error_kind": "ValueError"
1213
+ },
1214
+ {
1215
+ "reward_call": 26,
1216
+ "index": 1,
1217
+ "score": 0.814682099431488,
1218
+ "task_id": "bottle_cage_bike",
1219
+ "code_chars": 4275,
1220
+ "cadforge_build": 1.0,
1221
+ "cadforge_total": 0.7819789405076328,
1222
+ "has_fixture": true,
1223
+ "has_cadquery_import": true,
1224
+ "error_kind": "build_ok"
1225
+ },
1226
+ {
1227
+ "reward_call": 26,
1228
+ "index": 2,
1229
+ "score": -0.565,
1230
+ "task_id": "laptop_riser_vented",
1231
+ "code_chars": 4553,
1232
+ "cadforge_build": 0.0,
1233
+ "cadforge_total": -1.0,
1234
+ "has_fixture": true,
1235
+ "has_cadquery_import": true,
1236
+ "error_kind": "ValueError"
1237
+ },
1238
+ {
1239
+ "reward_call": 26,
1240
+ "index": 3,
1241
+ "score": -0.615,
1242
+ "task_id": "laptop_riser_vented",
1243
+ "code_chars": 4428,
1244
+ "cadforge_build": 0.0,
1245
+ "cadforge_total": -1.0,
1246
+ "has_fixture": true,
1247
+ "has_cadquery_import": true,
1248
+ "error_kind": "AttributeError"
1249
+ },
1250
+ {
1251
+ "reward_call": 27,
1252
+ "index": 0,
1253
+ "score": 0.6730844250208622,
1254
+ "task_id": "hinge_leaf_barrel",
1255
+ "code_chars": 5221,
1256
+ "cadforge_build": 1.0,
1257
+ "cadforge_total": 0.6153934412010144,
1258
+ "has_fixture": true,
1259
+ "has_cadquery_import": true,
1260
+ "error_kind": "build_ok"
1261
+ },
1262
+ {
1263
+ "reward_call": 27,
1264
+ "index": 1,
1265
+ "score": -0.8699999999999999,
1266
+ "task_id": "hinge_leaf_barrel",
1267
+ "code_chars": 5829,
1268
+ "cadforge_build": 0.0,
1269
+ "cadforge_total": -1.0,
1270
+ "has_fixture": false,
1271
+ "has_cadquery_import": true,
1272
+ "error_kind": "SyntaxError"
1273
+ },
1274
+ {
1275
+ "reward_call": 27,
1276
+ "index": 2,
1277
+ "score": -0.84,
1278
+ "task_id": "cable_clip_snap",
1279
+ "code_chars": 5789,
1280
+ "cadforge_build": 0.0,
1281
+ "cadforge_total": -1.0,
1282
+ "has_fixture": false,
1283
+ "has_cadquery_import": true,
1284
+ "error_kind": "SyntaxError"
1285
+ },
1286
+ {
1287
+ "reward_call": 27,
1288
+ "index": 3,
1289
+ "score": -0.74,
1290
+ "task_id": "cable_clip_snap",
1291
+ "code_chars": 5940,
1292
+ "cadforge_build": 0.0,
1293
+ "cadforge_total": -1.0,
1294
+ "has_fixture": false,
1295
+ "has_cadquery_import": true,
1296
+ "error_kind": "ValueError"
1297
+ },
1298
+ {
1299
+ "reward_call": 28,
1300
+ "index": 0,
1301
+ "score": -0.84,
1302
+ "task_id": "bench_vise_simplified",
1303
+ "code_chars": 5124,
1304
+ "cadforge_build": 0.0,
1305
+ "cadforge_total": -1.0,
1306
+ "has_fixture": false,
1307
+ "has_cadquery_import": true,
1308
+ "error_kind": "SyntaxError"
1309
+ },
1310
+ {
1311
+ "reward_call": 28,
1312
+ "index": 1,
1313
+ "score": -0.84,
1314
+ "task_id": "bench_vise_simplified",
1315
+ "code_chars": 5778,
1316
+ "cadforge_build": 0.0,
1317
+ "cadforge_total": -1.0,
1318
+ "has_fixture": false,
1319
+ "has_cadquery_import": true,
1320
+ "error_kind": "SyntaxError"
1321
+ },
1322
+ {
1323
+ "reward_call": 28,
1324
+ "index": 2,
1325
+ "score": -0.84,
1326
+ "task_id": "folding_step_stool",
1327
+ "code_chars": 5592,
1328
+ "cadforge_build": 0.0,
1329
+ "cadforge_total": -1.0,
1330
+ "has_fixture": false,
1331
+ "has_cadquery_import": true,
1332
+ "error_kind": "SyntaxError"
1333
+ },
1334
+ {
1335
+ "reward_call": 28,
1336
+ "index": 3,
1337
+ "score": -0.84,
1338
+ "task_id": "folding_step_stool",
1339
+ "code_chars": 5254,
1340
+ "cadforge_build": 0.0,
1341
+ "cadforge_total": -1.0,
1342
+ "has_fixture": false,
1343
+ "has_cadquery_import": true,
1344
+ "error_kind": "SyntaxError"
1345
+ },
1346
+ {
1347
+ "reward_call": 29,
1348
+ "index": 0,
1349
+ "score": 0.7325234226068589,
1350
+ "task_id": "drawer_handle_two_bolt",
1351
+ "code_chars": 2445,
1352
+ "cadforge_build": 1.0,
1353
+ "cadforge_total": 0.6853216736551281,
1354
+ "has_fixture": true,
1355
+ "has_cadquery_import": true,
1356
+ "error_kind": "build_ok"
1357
+ },
1358
+ {
1359
+ "reward_call": 29,
1360
+ "index": 1,
1361
+ "score": -0.595,
1362
+ "task_id": "drawer_handle_two_bolt",
1363
+ "code_chars": 3497,
1364
+ "cadforge_build": 0.0,
1365
+ "cadforge_total": -1.0,
1366
+ "has_fixture": true,
1367
+ "has_cadquery_import": true,
1368
+ "error_kind": "TypeError"
1369
+ },
1370
+ {
1371
+ "reward_call": 29,
1372
+ "index": 2,
1373
+ "score": -0.6649999999999999,
1374
+ "task_id": "shaft_torque_clamp_120nm",
1375
+ "code_chars": 5372,
1376
+ "cadforge_build": 0.0,
1377
+ "cadforge_total": -1.0,
1378
+ "has_fixture": true,
1379
+ "has_cadquery_import": true,
1380
+ "error_kind": "SyntaxError"
1381
+ },
1382
+ {
1383
+ "reward_call": 29,
1384
+ "index": 3,
1385
+ "score": -0.645,
1386
+ "task_id": "shaft_torque_clamp_120nm",
1387
+ "code_chars": 3407,
1388
+ "cadforge_build": 0.0,
1389
+ "cadforge_total": -1.0,
1390
+ "has_fixture": true,
1391
+ "has_cadquery_import": true,
1392
+ "error_kind": "NameError"
1393
+ },
1394
+ {
1395
+ "reward_call": 30,
1396
+ "index": 0,
1397
+ "score": -0.615,
1398
+ "task_id": "lightweight_truss_support_250n",
1399
+ "code_chars": 4762,
1400
+ "cadforge_build": 0.0,
1401
+ "cadforge_total": -1.0,
1402
+ "has_fixture": true,
1403
+ "has_cadquery_import": true,
1404
+ "error_kind": "AttributeError"
1405
+ },
1406
+ {
1407
+ "reward_call": 30,
1408
+ "index": 1,
1409
+ "score": -0.84,
1410
+ "task_id": "lightweight_truss_support_250n",
1411
+ "code_chars": 5053,
1412
+ "cadforge_build": 0.0,
1413
+ "cadforge_total": -1.0,
1414
+ "has_fixture": false,
1415
+ "has_cadquery_import": true,
1416
+ "error_kind": "SyntaxError"
1417
+ },
1418
+ {
1419
+ "reward_call": 30,
1420
+ "index": 2,
1421
+ "score": -0.6649999999999999,
1422
+ "task_id": "shelf_bracket_triangular_200n",
1423
+ "code_chars": 3097,
1424
+ "cadforge_build": 0.0,
1425
+ "cadforge_total": -1.0,
1426
+ "has_fixture": true,
1427
+ "has_cadquery_import": true,
1428
+ "error_kind": "SyntaxError"
1429
+ },
1430
+ {
1431
+ "reward_call": 30,
1432
+ "index": 3,
1433
+ "score": 0.8792378476009259,
1434
+ "task_id": "shelf_bracket_triangular_200n",
1435
+ "code_chars": 2662,
1436
+ "cadforge_build": 1.0,
1437
+ "cadforge_total": 0.8579268795305011,
1438
+ "has_fixture": true,
1439
+ "has_cadquery_import": true,
1440
+ "error_kind": "build_ok"
1441
+ },
1442
+ {
1443
+ "reward_call": 31,
1444
+ "index": 0,
1445
+ "score": -0.6649999999999999,
1446
+ "task_id": "desk_lamp_arm_joint",
1447
+ "code_chars": 6152,
1448
+ "cadforge_build": 0.0,
1449
+ "cadforge_total": -1.0,
1450
+ "has_fixture": true,
1451
+ "has_cadquery_import": true,
1452
+ "error_kind": "SyntaxError"
1453
+ },
1454
+ {
1455
+ "reward_call": 31,
1456
+ "index": 1,
1457
+ "score": -0.565,
1458
+ "task_id": "desk_lamp_arm_joint",
1459
+ "code_chars": 5784,
1460
+ "cadforge_build": 0.0,
1461
+ "cadforge_total": -1.0,
1462
+ "has_fixture": true,
1463
+ "has_cadquery_import": true,
1464
+ "error_kind": "TypeError"
1465
+ },
1466
+ {
1467
+ "reward_call": 31,
1468
+ "index": 2,
1469
+ "score": -0.565,
1470
+ "task_id": "robot_gripper_two_finger",
1471
+ "code_chars": 3974,
1472
+ "cadforge_build": 0.0,
1473
+ "cadforge_total": -1.0,
1474
+ "has_fixture": true,
1475
+ "has_cadquery_import": true,
1476
+ "error_kind": "TypeError"
1477
+ },
1478
+ {
1479
+ "reward_call": 31,
1480
+ "index": 3,
1481
+ "score": 0.8011607142857143,
1482
+ "task_id": "robot_gripper_two_finger",
1483
+ "code_chars": 4552,
1484
+ "cadforge_build": 1.0,
1485
+ "cadforge_total": 0.7660714285714286,
1486
+ "has_fixture": true,
1487
+ "has_cadquery_import": true,
1488
+ "error_kind": "build_ok"
1489
+ },
1490
+ {
1491
+ "reward_call": 32,
1492
+ "index": 0,
1493
+ "score": -0.595,
1494
+ "task_id": "pipe_clamp_saddle",
1495
+ "code_chars": 3251,
1496
+ "cadforge_build": 0.0,
1497
+ "cadforge_total": -1.0,
1498
+ "has_fixture": true,
1499
+ "has_cadquery_import": true,
1500
+ "error_kind": "StdFail"
1501
+ },
1502
+ {
1503
+ "reward_call": 32,
1504
+ "index": 1,
1505
+ "score": -0.84,
1506
+ "task_id": "pipe_clamp_saddle",
1507
+ "code_chars": 6271,
1508
+ "cadforge_build": 0.0,
1509
+ "cadforge_total": -1.0,
1510
+ "has_fixture": false,
1511
+ "has_cadquery_import": true,
1512
+ "error_kind": "SyntaxError"
1513
+ },
1514
+ {
1515
+ "reward_call": 32,
1516
+ "index": 2,
1517
+ "score": -0.61,
1518
+ "task_id": "wall_j_hook_120n",
1519
+ "code_chars": 4150,
1520
+ "cadforge_build": 0.0,
1521
+ "cadforge_total": -1.0,
1522
+ "has_fixture": true,
1523
+ "has_cadquery_import": true,
1524
+ "error_kind": "unknown"
1525
+ },
1526
+ {
1527
+ "reward_call": 32,
1528
+ "index": 3,
1529
+ "score": -0.565,
1530
+ "task_id": "wall_j_hook_120n",
1531
+ "code_chars": 3835,
1532
+ "cadforge_build": 0.0,
1533
+ "cadforge_total": -1.0,
1534
+ "has_fixture": true,
1535
+ "has_cadquery_import": true,
1536
+ "error_kind": "ValueError"
1537
+ },
1538
+ {
1539
+ "reward_call": 33,
1540
+ "index": 0,
1541
+ "score": -0.565,
1542
+ "task_id": "laptop_riser_vented",
1543
+ "code_chars": 4353,
1544
+ "cadforge_build": 0.0,
1545
+ "cadforge_total": -1.0,
1546
+ "has_fixture": true,
1547
+ "has_cadquery_import": true,
1548
+ "error_kind": "ValueError"
1549
+ },
1550
+ {
1551
+ "reward_call": 33,
1552
+ "index": 1,
1553
+ "score": 0.875122259801786,
1554
+ "task_id": "laptop_riser_vented",
1555
+ "code_chars": 3799,
1556
+ "cadforge_build": 1.0,
1557
+ "cadforge_total": 0.853085011531513,
1558
+ "has_fixture": true,
1559
+ "has_cadquery_import": true,
1560
+ "error_kind": "build_ok"
1561
+ },
1562
+ {
1563
+ "reward_call": 33,
1564
+ "index": 2,
1565
+ "score": -0.84,
1566
+ "task_id": "bench_vise_simplified",
1567
+ "code_chars": 5867,
1568
+ "cadforge_build": 0.0,
1569
+ "cadforge_total": -1.0,
1570
+ "has_fixture": false,
1571
+ "has_cadquery_import": true,
1572
+ "error_kind": "SyntaxError"
1573
+ },
1574
+ {
1575
+ "reward_call": 33,
1576
+ "index": 3,
1577
+ "score": -0.84,
1578
+ "task_id": "bench_vise_simplified",
1579
+ "code_chars": 5434,
1580
+ "cadforge_build": 0.0,
1581
+ "cadforge_total": -1.0,
1582
+ "has_fixture": false,
1583
+ "has_cadquery_import": true,
1584
+ "error_kind": "SyntaxError"
1585
+ },
1586
+ {
1587
+ "reward_call": 34,
1588
+ "index": 0,
1589
+ "score": 0.6869166666666668,
1590
+ "task_id": "table_six_leg_500n",
1591
+ "code_chars": 3276,
1592
+ "cadforge_build": 1.0,
1593
+ "cadforge_total": 0.6316666666666668,
1594
+ "has_fixture": true,
1595
+ "has_cadquery_import": true,
1596
+ "error_kind": "build_ok"
1597
+ },
1598
+ {
1599
+ "reward_call": 34,
1600
+ "index": 1,
1601
+ "score": -0.615,
1602
+ "task_id": "table_six_leg_500n",
1603
+ "code_chars": 3561,
1604
+ "cadforge_build": 0.0,
1605
+ "cadforge_total": -1.0,
1606
+ "has_fixture": true,
1607
+ "has_cadquery_import": true,
1608
+ "error_kind": "NameError"
1609
+ },
1610
+ {
1611
+ "reward_call": 34,
1612
+ "index": 2,
1613
+ "score": -0.6649999999999999,
1614
+ "task_id": "lightweight_truss_support_250n",
1615
+ "code_chars": 5589,
1616
+ "cadforge_build": 0.0,
1617
+ "cadforge_total": -1.0,
1618
+ "has_fixture": true,
1619
+ "has_cadquery_import": true,
1620
+ "error_kind": "SyntaxError"
1621
+ },
1622
+ {
1623
+ "reward_call": 34,
1624
+ "index": 3,
1625
+ "score": -0.595,
1626
+ "task_id": "lightweight_truss_support_250n",
1627
+ "code_chars": 4236,
1628
+ "cadforge_build": 0.0,
1629
+ "cadforge_total": -1.0,
1630
+ "has_fixture": true,
1631
+ "has_cadquery_import": true,
1632
+ "error_kind": "TypeError"
1633
+ },
1634
+ {
1635
+ "reward_call": 35,
1636
+ "index": 0,
1637
+ "score": 0.8842360448284298,
1638
+ "task_id": "phone_stand_adjustable_hinge",
1639
+ "code_chars": 3470,
1640
+ "cadforge_build": 1.0,
1641
+ "cadforge_total": 0.8638071115628585,
1642
+ "has_fixture": true,
1643
+ "has_cadquery_import": true,
1644
+ "error_kind": "build_ok"
1645
+ },
1646
+ {
1647
+ "reward_call": 35,
1648
+ "index": 1,
1649
+ "score": 0.9193229814943313,
1650
+ "task_id": "phone_stand_adjustable_hinge",
1651
+ "code_chars": 5465,
1652
+ "cadforge_build": 1.0,
1653
+ "cadforge_total": 0.9050858605815663,
1654
+ "has_fixture": true,
1655
+ "has_cadquery_import": true,
1656
+ "error_kind": "build_ok"
1657
+ },
1658
+ {
1659
+ "reward_call": 35,
1660
+ "index": 2,
1661
+ "score": -0.565,
1662
+ "task_id": "bike_accessory_mount_120n",
1663
+ "code_chars": 5764,
1664
+ "cadforge_build": 0.0,
1665
+ "cadforge_total": -1.0,
1666
+ "has_fixture": true,
1667
+ "has_cadquery_import": true,
1668
+ "error_kind": "TypeError"
1669
+ },
1670
+ {
1671
+ "reward_call": 35,
1672
+ "index": 3,
1673
+ "score": -0.615,
1674
+ "task_id": "bike_accessory_mount_120n",
1675
+ "code_chars": 5731,
1676
+ "cadforge_build": 0.0,
1677
+ "cadforge_total": -1.0,
1678
+ "has_fixture": true,
1679
+ "has_cadquery_import": true,
1680
+ "error_kind": "NameError"
1681
+ },
1682
+ {
1683
+ "reward_call": 36,
1684
+ "index": 0,
1685
+ "score": -0.615,
1686
+ "task_id": "drawer_handle_two_bolt",
1687
+ "code_chars": 4283,
1688
+ "cadforge_build": 0.0,
1689
+ "cadforge_total": -1.0,
1690
+ "has_fixture": true,
1691
+ "has_cadquery_import": true,
1692
+ "error_kind": "NameError"
1693
+ },
1694
+ {
1695
+ "reward_call": 36,
1696
+ "index": 1,
1697
+ "score": 0.6973552631578949,
1698
+ "task_id": "drawer_handle_two_bolt",
1699
+ "code_chars": 2996,
1700
+ "cadforge_build": 1.0,
1701
+ "cadforge_total": 0.6439473684210527,
1702
+ "has_fixture": true,
1703
+ "has_cadquery_import": true,
1704
+ "error_kind": "build_ok"
1705
+ },
1706
+ {
1707
+ "reward_call": 36,
1708
+ "index": 2,
1709
+ "score": -0.84,
1710
+ "task_id": "ergonomic_curvy_chair_1000n",
1711
+ "code_chars": 4689,
1712
+ "cadforge_build": 0.0,
1713
+ "cadforge_total": -1.0,
1714
+ "has_fixture": false,
1715
+ "has_cadquery_import": true,
1716
+ "error_kind": "SyntaxError"
1717
+ },
1718
+ {
1719
+ "reward_call": 36,
1720
+ "index": 3,
1721
+ "score": -0.84,
1722
+ "task_id": "ergonomic_curvy_chair_1000n",
1723
+ "code_chars": 4988,
1724
+ "cadforge_build": 0.0,
1725
+ "cadforge_total": -1.0,
1726
+ "has_fixture": false,
1727
+ "has_cadquery_import": true,
1728
+ "error_kind": "SyntaxError"
1729
+ },
1730
+ {
1731
+ "reward_call": 37,
1732
+ "index": 0,
1733
+ "score": 0.7053947435579438,
1734
+ "task_id": "shelf_bracket_triangular_200n",
1735
+ "code_chars": 2517,
1736
+ "cadforge_build": 1.0,
1737
+ "cadforge_total": 0.6534055806564045,
1738
+ "has_fixture": true,
1739
+ "has_cadquery_import": true,
1740
+ "error_kind": "build_ok"
1741
+ },
1742
+ {
1743
+ "reward_call": 37,
1744
+ "index": 1,
1745
+ "score": 0.7518514895513351,
1746
+ "task_id": "shelf_bracket_triangular_200n",
1747
+ "code_chars": 3494,
1748
+ "cadforge_build": 1.0,
1749
+ "cadforge_total": 0.7080605759427472,
1750
+ "has_fixture": true,
1751
+ "has_cadquery_import": true,
1752
+ "error_kind": "build_ok"
1753
+ },
1754
+ {
1755
+ "reward_call": 37,
1756
+ "index": 2,
1757
+ "score": -0.6649999999999999,
1758
+ "task_id": "micro_drone_frame",
1759
+ "code_chars": 4736,
1760
+ "cadforge_build": 0.0,
1761
+ "cadforge_total": -1.0,
1762
+ "has_fixture": true,
1763
+ "has_cadquery_import": true,
1764
+ "error_kind": "SyntaxError"
1765
+ },
1766
+ {
1767
+ "reward_call": 37,
1768
+ "index": 3,
1769
+ "score": -0.565,
1770
+ "task_id": "micro_drone_frame",
1771
+ "code_chars": 5674,
1772
+ "cadforge_build": 0.0,
1773
+ "cadforge_total": -1.0,
1774
+ "has_fixture": true,
1775
+ "has_cadquery_import": true,
1776
+ "error_kind": "TypeError"
1777
+ },
1778
+ {
1779
+ "reward_call": 38,
1780
+ "index": 0,
1781
+ "score": -0.79,
1782
+ "task_id": "gearbox_housing_split",
1783
+ "code_chars": 6152,
1784
+ "cadforge_build": 0.0,
1785
+ "cadforge_total": -1.0,
1786
+ "has_fixture": false,
1787
+ "has_cadquery_import": true,
1788
+ "error_kind": "NameError"
1789
+ },
1790
+ {
1791
+ "reward_call": 38,
1792
+ "index": 1,
1793
+ "score": -0.84,
1794
+ "task_id": "gearbox_housing_split",
1795
+ "code_chars": 5504,
1796
+ "cadforge_build": 0.0,
1797
+ "cadforge_total": -1.0,
1798
+ "has_fixture": false,
1799
+ "has_cadquery_import": true,
1800
+ "error_kind": "SyntaxError"
1801
+ },
1802
+ {
1803
+ "reward_call": 38,
1804
+ "index": 2,
1805
+ "score": -0.79,
1806
+ "task_id": "cable_clip_snap",
1807
+ "code_chars": 5898,
1808
+ "cadforge_build": 0.0,
1809
+ "cadforge_total": -1.0,
1810
+ "has_fixture": false,
1811
+ "has_cadquery_import": true,
1812
+ "error_kind": "AttributeError"
1813
+ },
1814
+ {
1815
+ "reward_call": 38,
1816
+ "index": 3,
1817
+ "score": -0.6649999999999999,
1818
+ "task_id": "cable_clip_snap",
1819
+ "code_chars": 5345,
1820
+ "cadforge_build": 0.0,
1821
+ "cadforge_total": -1.0,
1822
+ "has_fixture": true,
1823
+ "has_cadquery_import": true,
1824
+ "error_kind": "SyntaxError"
1825
+ },
1826
+ {
1827
+ "reward_call": 39,
1828
+ "index": 0,
1829
+ "score": 0.6642500000000001,
1830
+ "task_id": "bottle_cage_bike",
1831
+ "code_chars": 4802,
1832
+ "cadforge_build": 1.0,
1833
+ "cadforge_total": 0.6050000000000001,
1834
+ "has_fixture": true,
1835
+ "has_cadquery_import": true,
1836
+ "error_kind": "build_ok"
1837
+ },
1838
+ {
1839
+ "reward_call": 39,
1840
+ "index": 1,
1841
+ "score": 0.8274949403962913,
1842
+ "task_id": "bottle_cage_bike",
1843
+ "code_chars": 4829,
1844
+ "cadforge_build": 1.0,
1845
+ "cadforge_total": 0.7970528710544604,
1846
+ "has_fixture": true,
1847
+ "has_cadquery_import": true,
1848
+ "error_kind": "build_ok"
1849
+ },
1850
+ {
1851
+ "reward_call": 39,
1852
+ "index": 2,
1853
+ "score": -0.74,
1854
+ "task_id": "folding_step_stool",
1855
+ "code_chars": 5732,
1856
+ "cadforge_build": 0.0,
1857
+ "cadforge_total": -1.0,
1858
+ "has_fixture": false,
1859
+ "has_cadquery_import": true,
1860
+ "error_kind": "ValueError"
1861
+ },
1862
+ {
1863
+ "reward_call": 39,
1864
+ "index": 3,
1865
+ "score": -0.84,
1866
+ "task_id": "folding_step_stool",
1867
+ "code_chars": 5252,
1868
+ "cadforge_build": 0.0,
1869
+ "cadforge_total": -1.0,
1870
+ "has_fixture": false,
1871
+ "has_cadquery_import": true,
1872
+ "error_kind": "SyntaxError"
1873
+ },
1874
+ {
1875
+ "reward_call": 40,
1876
+ "index": 0,
1877
+ "score": -0.565,
1878
+ "task_id": "hinge_leaf_barrel",
1879
+ "code_chars": 3500,
1880
+ "cadforge_build": 0.0,
1881
+ "cadforge_total": -1.0,
1882
+ "has_fixture": true,
1883
+ "has_cadquery_import": true,
1884
+ "error_kind": "TypeError"
1885
+ },
1886
+ {
1887
+ "reward_call": 40,
1888
+ "index": 1,
1889
+ "score": -0.695,
1890
+ "task_id": "hinge_leaf_barrel",
1891
+ "code_chars": 6200,
1892
+ "cadforge_build": 0.0,
1893
+ "cadforge_total": -1.0,
1894
+ "has_fixture": true,
1895
+ "has_cadquery_import": true,
1896
+ "error_kind": "SyntaxError"
1897
+ },
1898
+ {
1899
+ "reward_call": 40,
1900
+ "index": 2,
1901
+ "score": 0.7114237349456013,
1902
+ "task_id": "shaft_torque_clamp_120nm",
1903
+ "code_chars": 4587,
1904
+ "cadforge_build": 1.0,
1905
+ "cadforge_total": 0.6604985117007074,
1906
+ "has_fixture": true,
1907
+ "has_cadquery_import": true,
1908
+ "error_kind": "build_ok"
1909
+ },
1910
+ {
1911
+ "reward_call": 40,
1912
+ "index": 3,
1913
+ "score": 0.7068521044496177,
1914
+ "task_id": "shaft_torque_clamp_120nm",
1915
+ "code_chars": 3359,
1916
+ "cadforge_build": 1.0,
1917
+ "cadforge_total": 0.6551201228819031,
1918
+ "has_fixture": true,
1919
+ "has_cadquery_import": true,
1920
+ "error_kind": "build_ok"
1921
+ },
1922
+ {
1923
+ "reward_call": 41,
1924
+ "index": 0,
1925
+ "score": -0.565,
1926
+ "task_id": "cable_clip_snap",
1927
+ "code_chars": 3141,
1928
+ "cadforge_build": 0.0,
1929
+ "cadforge_total": -1.0,
1930
+ "has_fixture": true,
1931
+ "has_cadquery_import": true,
1932
+ "error_kind": "BRep_API"
1933
+ },
1934
+ {
1935
+ "reward_call": 41,
1936
+ "index": 1,
1937
+ "score": -0.565,
1938
+ "task_id": "cable_clip_snap",
1939
+ "code_chars": 3957,
1940
+ "cadforge_build": 0.0,
1941
+ "cadforge_total": -1.0,
1942
+ "has_fixture": true,
1943
+ "has_cadquery_import": true,
1944
+ "error_kind": "unknown"
1945
+ },
1946
+ {
1947
+ "reward_call": 41,
1948
+ "index": 2,
1949
+ "score": 0.871946352186128,
1950
+ "task_id": "pipe_clamp_saddle",
1951
+ "code_chars": 2954,
1952
+ "cadforge_build": 1.0,
1953
+ "cadforge_total": 0.8493486496307389,
1954
+ "has_fixture": true,
1955
+ "has_cadquery_import": true,
1956
+ "error_kind": "build_ok"
1957
+ },
1958
+ {
1959
+ "reward_call": 41,
1960
+ "index": 3,
1961
+ "score": -0.6649999999999999,
1962
+ "task_id": "pipe_clamp_saddle",
1963
+ "code_chars": 2681,
1964
+ "cadforge_build": 0.0,
1965
+ "cadforge_total": -1.0,
1966
+ "has_fixture": true,
1967
+ "has_cadquery_import": true,
1968
+ "error_kind": "SyntaxError"
1969
+ },
1970
+ {
1971
+ "reward_call": 42,
1972
+ "index": 0,
1973
+ "score": -0.615,
1974
+ "task_id": "laptop_riser_vented",
1975
+ "code_chars": 4234,
1976
+ "cadforge_build": 0.0,
1977
+ "cadforge_total": -1.0,
1978
+ "has_fixture": true,
1979
+ "has_cadquery_import": true,
1980
+ "error_kind": "NameError"
1981
+ },
1982
+ {
1983
+ "reward_call": 42,
1984
+ "index": 1,
1985
+ "score": -0.6649999999999999,
1986
+ "task_id": "laptop_riser_vented",
1987
+ "code_chars": 3876,
1988
+ "cadforge_build": 0.0,
1989
+ "cadforge_total": -1.0,
1990
+ "has_fixture": true,
1991
+ "has_cadquery_import": true,
1992
+ "error_kind": "SyntaxError"
1993
+ },
1994
+ {
1995
+ "reward_call": 42,
1996
+ "index": 2,
1997
+ "score": -0.615,
1998
+ "task_id": "hinge_leaf_barrel",
1999
+ "code_chars": 4904,
2000
+ "cadforge_build": 0.0,
2001
+ "cadforge_total": -1.0,
2002
+ "has_fixture": true,
2003
+ "has_cadquery_import": true,
2004
+ "error_kind": "AttributeError"
2005
+ },
2006
+ {
2007
+ "reward_call": 42,
2008
+ "index": 3,
2009
+ "score": -0.565,
2010
+ "task_id": "hinge_leaf_barrel",
2011
+ "code_chars": 4744,
2012
+ "cadforge_build": 0.0,
2013
+ "cadforge_total": -1.0,
2014
+ "has_fixture": true,
2015
+ "has_cadquery_import": true,
2016
+ "error_kind": "ValueError"
2017
+ },
2018
+ {
2019
+ "reward_call": 43,
2020
+ "index": 0,
2021
+ "score": -0.6649999999999999,
2022
+ "task_id": "bottle_cage_bike",
2023
+ "code_chars": 6217,
2024
+ "cadforge_build": 0.0,
2025
+ "cadforge_total": -1.0,
2026
+ "has_fixture": true,
2027
+ "has_cadquery_import": true,
2028
+ "error_kind": "SyntaxError"
2029
+ },
2030
+ {
2031
+ "reward_call": 43,
2032
+ "index": 1,
2033
+ "score": -0.565,
2034
+ "task_id": "bottle_cage_bike",
2035
+ "code_chars": 4560,
2036
+ "cadforge_build": 0.0,
2037
+ "cadforge_total": -1.0,
2038
+ "has_fixture": true,
2039
+ "has_cadquery_import": true,
2040
+ "error_kind": "ValueError"
2041
+ },
2042
+ {
2043
+ "reward_call": 43,
2044
+ "index": 2,
2045
+ "score": -0.84,
2046
+ "task_id": "ergonomic_curvy_chair_1000n",
2047
+ "code_chars": 5147,
2048
+ "cadforge_build": 0.0,
2049
+ "cadforge_total": -1.0,
2050
+ "has_fixture": false,
2051
+ "has_cadquery_import": true,
2052
+ "error_kind": "SyntaxError"
2053
+ },
2054
+ {
2055
+ "reward_call": 43,
2056
+ "index": 3,
2057
+ "score": -0.84,
2058
+ "task_id": "ergonomic_curvy_chair_1000n",
2059
+ "code_chars": 4709,
2060
+ "cadforge_build": 0.0,
2061
+ "cadforge_total": -1.0,
2062
+ "has_fixture": false,
2063
+ "has_cadquery_import": true,
2064
+ "error_kind": "SyntaxError"
2065
+ },
2066
+ {
2067
+ "reward_call": 44,
2068
+ "index": 0,
2069
+ "score": -0.615,
2070
+ "task_id": "micro_drone_frame",
2071
+ "code_chars": 4444,
2072
+ "cadforge_build": 0.0,
2073
+ "cadforge_total": -1.0,
2074
+ "has_fixture": true,
2075
+ "has_cadquery_import": true,
2076
+ "error_kind": "AttributeError"
2077
+ },
2078
+ {
2079
+ "reward_call": 44,
2080
+ "index": 1,
2081
+ "score": -0.79,
2082
+ "task_id": "micro_drone_frame",
2083
+ "code_chars": 5540,
2084
+ "cadforge_build": 0.0,
2085
+ "cadforge_total": -1.0,
2086
+ "has_fixture": false,
2087
+ "has_cadquery_import": true,
2088
+ "error_kind": "NameError"
2089
+ },
2090
+ {
2091
+ "reward_call": 44,
2092
+ "index": 2,
2093
+ "score": -0.6649999999999999,
2094
+ "task_id": "table_six_leg_500n",
2095
+ "code_chars": 4577,
2096
+ "cadforge_build": 0.0,
2097
+ "cadforge_total": -1.0,
2098
+ "has_fixture": true,
2099
+ "has_cadquery_import": true,
2100
+ "error_kind": "SyntaxError"
2101
+ },
2102
+ {
2103
+ "reward_call": 44,
2104
+ "index": 3,
2105
+ "score": 0.8346430625994663,
2106
+ "task_id": "table_six_leg_500n",
2107
+ "code_chars": 3110,
2108
+ "cadforge_build": 1.0,
2109
+ "cadforge_total": 0.8054624265876075,
2110
+ "has_fixture": true,
2111
+ "has_cadquery_import": true,
2112
+ "error_kind": "build_ok"
2113
+ },
2114
+ {
2115
+ "reward_call": 45,
2116
+ "index": 0,
2117
+ "score": -0.84,
2118
+ "task_id": "folding_step_stool",
2119
+ "code_chars": 4872,
2120
+ "cadforge_build": 0.0,
2121
+ "cadforge_total": -1.0,
2122
+ "has_fixture": false,
2123
+ "has_cadquery_import": true,
2124
+ "error_kind": "SyntaxError"
2125
+ },
2126
+ {
2127
+ "reward_call": 45,
2128
+ "index": 1,
2129
+ "score": -0.79,
2130
+ "task_id": "folding_step_stool",
2131
+ "code_chars": 5138,
2132
+ "cadforge_build": 0.0,
2133
+ "cadforge_total": -1.0,
2134
+ "has_fixture": false,
2135
+ "has_cadquery_import": true,
2136
+ "error_kind": "AttributeError"
2137
+ },
2138
+ {
2139
+ "reward_call": 45,
2140
+ "index": 2,
2141
+ "score": 0.7470980837605303,
2142
+ "task_id": "desk_lamp_arm_joint",
2143
+ "code_chars": 5844,
2144
+ "cadforge_build": 1.0,
2145
+ "cadforge_total": 0.702468333835918,
2146
+ "has_fixture": true,
2147
+ "has_cadquery_import": true,
2148
+ "error_kind": "build_ok"
2149
+ },
2150
+ {
2151
+ "reward_call": 45,
2152
+ "index": 3,
2153
+ "score": 0.6401666666666668,
2154
+ "task_id": "desk_lamp_arm_joint",
2155
+ "code_chars": 4613,
2156
+ "cadforge_build": 1.0,
2157
+ "cadforge_total": 0.5766666666666668,
2158
+ "has_fixture": true,
2159
+ "has_cadquery_import": true,
2160
+ "error_kind": "build_ok"
2161
+ },
2162
+ {
2163
+ "reward_call": 46,
2164
+ "index": 0,
2165
+ "score": -0.61,
2166
+ "task_id": "wall_j_hook_120n",
2167
+ "code_chars": 3407,
2168
+ "cadforge_build": 0.0,
2169
+ "cadforge_total": -1.0,
2170
+ "has_fixture": true,
2171
+ "has_cadquery_import": true,
2172
+ "error_kind": "TypeError"
2173
+ },
2174
+ {
2175
+ "reward_call": 46,
2176
+ "index": 1,
2177
+ "score": -0.565,
2178
+ "task_id": "wall_j_hook_120n",
2179
+ "code_chars": 3003,
2180
+ "cadforge_build": 0.0,
2181
+ "cadforge_total": -1.0,
2182
+ "has_fixture": true,
2183
+ "has_cadquery_import": true,
2184
+ "error_kind": "ValueError"
2185
+ },
2186
+ {
2187
+ "reward_call": 46,
2188
+ "index": 2,
2189
+ "score": 0.8369582933584849,
2190
+ "task_id": "lightweight_truss_support_250n",
2191
+ "code_chars": 4804,
2192
+ "cadforge_build": 1.0,
2193
+ "cadforge_total": 0.8081862274805705,
2194
+ "has_fixture": true,
2195
+ "has_cadquery_import": true,
2196
+ "error_kind": "build_ok"
2197
+ },
2198
+ {
2199
+ "reward_call": 46,
2200
+ "index": 3,
2201
+ "score": -0.84,
2202
+ "task_id": "lightweight_truss_support_250n",
2203
+ "code_chars": 4454,
2204
+ "cadforge_build": 0.0,
2205
+ "cadforge_total": -1.0,
2206
+ "has_fixture": false,
2207
+ "has_cadquery_import": true,
2208
+ "error_kind": "SyntaxError"
2209
+ },
2210
+ {
2211
+ "reward_call": 47,
2212
+ "index": 0,
2213
+ "score": -0.615,
2214
+ "task_id": "robot_gripper_two_finger",
2215
+ "code_chars": 5250,
2216
+ "cadforge_build": 0.0,
2217
+ "cadforge_total": -1.0,
2218
+ "has_fixture": true,
2219
+ "has_cadquery_import": true,
2220
+ "error_kind": "AttributeError"
2221
+ },
2222
+ {
2223
+ "reward_call": 47,
2224
+ "index": 1,
2225
+ "score": 0.6997138900547093,
2226
+ "task_id": "robot_gripper_two_finger",
2227
+ "code_chars": 3424,
2228
+ "cadforge_build": 1.0,
2229
+ "cadforge_total": 0.6467222235937756,
2230
+ "has_fixture": true,
2231
+ "has_cadquery_import": true,
2232
+ "error_kind": "build_ok"
2233
+ },
2234
+ {
2235
+ "reward_call": 47,
2236
+ "index": 2,
2237
+ "score": -0.84,
2238
+ "task_id": "gearbox_housing_split",
2239
+ "code_chars": 6428,
2240
+ "cadforge_build": 0.0,
2241
+ "cadforge_total": -1.0,
2242
+ "has_fixture": false,
2243
+ "has_cadquery_import": true,
2244
+ "error_kind": "SyntaxError"
2245
+ },
2246
+ {
2247
+ "reward_call": 47,
2248
+ "index": 3,
2249
+ "score": -0.84,
2250
+ "task_id": "gearbox_housing_split",
2251
+ "code_chars": 5680,
2252
+ "cadforge_build": 0.0,
2253
+ "cadforge_total": -1.0,
2254
+ "has_fixture": false,
2255
+ "has_cadquery_import": true,
2256
+ "error_kind": "SyntaxError"
2257
+ },
2258
+ {
2259
+ "reward_call": 48,
2260
+ "index": 0,
2261
+ "score": 0.6672099629296073,
2262
+ "task_id": "drawer_handle_two_bolt",
2263
+ "code_chars": 3586,
2264
+ "cadforge_build": 1.0,
2265
+ "cadforge_total": 0.6084823093289498,
2266
+ "has_fixture": true,
2267
+ "has_cadquery_import": true,
2268
+ "error_kind": "build_ok"
2269
+ },
2270
+ {
2271
+ "reward_call": 48,
2272
+ "index": 1,
2273
+ "score": 0.6671490277904167,
2274
+ "task_id": "drawer_handle_two_bolt",
2275
+ "code_chars": 2160,
2276
+ "cadforge_build": 1.0,
2277
+ "cadforge_total": 0.6084106209299021,
2278
+ "has_fixture": true,
2279
+ "has_cadquery_import": true,
2280
+ "error_kind": "build_ok"
2281
+ },
2282
+ {
2283
+ "reward_call": 48,
2284
+ "index": 2,
2285
+ "score": 0.7806319796949592,
2286
+ "task_id": "shelf_bracket_triangular_200n",
2287
+ "code_chars": 2611,
2288
+ "cadforge_build": 1.0,
2289
+ "cadforge_total": 0.7419199761117168,
2290
+ "has_fixture": true,
2291
+ "has_cadquery_import": true,
2292
+ "error_kind": "build_ok"
2293
+ },
2294
+ {
2295
+ "reward_call": 48,
2296
+ "index": 3,
2297
+ "score": -0.615,
2298
+ "task_id": "shelf_bracket_triangular_200n",
2299
+ "code_chars": 2611,
2300
+ "cadforge_build": 0.0,
2301
+ "cadforge_total": -1.0,
2302
+ "has_fixture": true,
2303
+ "has_cadquery_import": true,
2304
+ "error_kind": "NameError"
2305
+ },
2306
+ {
2307
+ "reward_call": 49,
2308
+ "index": 0,
2309
+ "score": -0.6649999999999999,
2310
+ "task_id": "bike_accessory_mount_120n",
2311
+ "code_chars": 2505,
2312
+ "cadforge_build": 0.0,
2313
+ "cadforge_total": -1.0,
2314
+ "has_fixture": true,
2315
+ "has_cadquery_import": true,
2316
+ "error_kind": "SyntaxError"
2317
+ },
2318
+ {
2319
+ "reward_call": 49,
2320
+ "index": 1,
2321
+ "score": 0.7575464783499094,
2322
+ "task_id": "bike_accessory_mount_120n",
2323
+ "code_chars": 3589,
2324
+ "cadforge_build": 1.0,
2325
+ "cadforge_total": 0.7147605627645993,
2326
+ "has_fixture": true,
2327
+ "has_cadquery_import": true,
2328
+ "error_kind": "build_ok"
2329
+ },
2330
+ {
2331
+ "reward_call": 49,
2332
+ "index": 2,
2333
+ "score": -0.565,
2334
+ "task_id": "shaft_torque_clamp_120nm",
2335
+ "code_chars": 4131,
2336
+ "cadforge_build": 0.0,
2337
+ "cadforge_total": -1.0,
2338
+ "has_fixture": true,
2339
+ "has_cadquery_import": true,
2340
+ "error_kind": "ValueError"
2341
+ },
2342
+ {
2343
+ "reward_call": 49,
2344
+ "index": 3,
2345
+ "score": -0.565,
2346
+ "task_id": "shaft_torque_clamp_120nm",
2347
+ "code_chars": 5495,
2348
+ "cadforge_build": 0.0,
2349
+ "cadforge_total": -1.0,
2350
+ "has_fixture": true,
2351
+ "has_cadquery_import": true,
2352
+ "error_kind": "unknown"
2353
+ },
2354
+ {
2355
+ "reward_call": 50,
2356
+ "index": 0,
2357
+ "score": -0.84,
2358
+ "task_id": "bench_vise_simplified",
2359
+ "code_chars": 5814,
2360
+ "cadforge_build": 0.0,
2361
+ "cadforge_total": -1.0,
2362
+ "has_fixture": false,
2363
+ "has_cadquery_import": true,
2364
+ "error_kind": "SyntaxError"
2365
+ },
2366
+ {
2367
+ "reward_call": 50,
2368
+ "index": 1,
2369
+ "score": -0.6649999999999999,
2370
+ "task_id": "bench_vise_simplified",
2371
+ "code_chars": 4253,
2372
+ "cadforge_build": 0.0,
2373
+ "cadforge_total": -1.0,
2374
+ "has_fixture": true,
2375
+ "has_cadquery_import": true,
2376
+ "error_kind": "SyntaxError"
2377
+ },
2378
+ {
2379
+ "reward_call": 50,
2380
+ "index": 2,
2381
+ "score": -0.565,
2382
+ "task_id": "phone_stand_adjustable_hinge",
2383
+ "code_chars": 4290,
2384
+ "cadforge_build": 0.0,
2385
+ "cadforge_total": -1.0,
2386
+ "has_fixture": true,
2387
+ "has_cadquery_import": true,
2388
+ "error_kind": "unknown"
2389
+ },
2390
+ {
2391
+ "reward_call": 50,
2392
+ "index": 3,
2393
+ "score": 0.8716593817974053,
2394
+ "task_id": "phone_stand_adjustable_hinge",
2395
+ "code_chars": 5004,
2396
+ "cadforge_build": 1.0,
2397
+ "cadforge_total": 0.8490110374087122,
2398
+ "has_fixture": true,
2399
+ "has_cadquery_import": true,
2400
+ "error_kind": "build_ok"
2401
+ },
2402
+ {
2403
+ "reward_call": 51,
2404
+ "index": 0,
2405
+ "score": -0.645,
2406
+ "task_id": "drawer_handle_two_bolt",
2407
+ "code_chars": 3333,
2408
+ "cadforge_build": 0.0,
2409
+ "cadforge_total": -1.0,
2410
+ "has_fixture": true,
2411
+ "has_cadquery_import": true,
2412
+ "error_kind": "NameError"
2413
+ },
2414
+ {
2415
+ "reward_call": 51,
2416
+ "index": 1,
2417
+ "score": -0.565,
2418
+ "task_id": "drawer_handle_two_bolt",
2419
+ "code_chars": 4063,
2420
+ "cadforge_build": 0.0,
2421
+ "cadforge_total": -1.0,
2422
+ "has_fixture": true,
2423
+ "has_cadquery_import": true,
2424
+ "error_kind": "BRep_API"
2425
+ },
2426
+ {
2427
+ "reward_call": 51,
2428
+ "index": 2,
2429
+ "score": -0.615,
2430
+ "task_id": "shelf_bracket_triangular_200n",
2431
+ "code_chars": 3539,
2432
+ "cadforge_build": 0.0,
2433
+ "cadforge_total": -1.0,
2434
+ "has_fixture": true,
2435
+ "has_cadquery_import": true,
2436
+ "error_kind": "NameError"
2437
+ },
2438
+ {
2439
+ "reward_call": 51,
2440
+ "index": 3,
2441
+ "score": 0.7474324364269804,
2442
+ "task_id": "shelf_bracket_triangular_200n",
2443
+ "code_chars": 3226,
2444
+ "cadforge_build": 1.0,
2445
+ "cadforge_total": 0.7028616899140946,
2446
+ "has_fixture": true,
2447
+ "has_cadquery_import": true,
2448
+ "error_kind": "build_ok"
2449
+ },
2450
+ {
2451
+ "reward_call": 52,
2452
+ "index": 0,
2453
+ "score": -0.565,
2454
+ "task_id": "hinge_leaf_barrel",
2455
+ "code_chars": 4252,
2456
+ "cadforge_build": 0.0,
2457
+ "cadforge_total": -1.0,
2458
+ "has_fixture": true,
2459
+ "has_cadquery_import": true,
2460
+ "error_kind": "unknown"
2461
+ },
2462
+ {
2463
+ "reward_call": 52,
2464
+ "index": 1,
2465
+ "score": -0.595,
2466
+ "task_id": "hinge_leaf_barrel",
2467
+ "code_chars": 4008,
2468
+ "cadforge_build": 0.0,
2469
+ "cadforge_total": -1.0,
2470
+ "has_fixture": true,
2471
+ "has_cadquery_import": true,
2472
+ "error_kind": "ValueError"
2473
+ },
2474
+ {
2475
+ "reward_call": 52,
2476
+ "index": 2,
2477
+ "score": -0.565,
2478
+ "task_id": "desk_lamp_arm_joint",
2479
+ "code_chars": 5156,
2480
+ "cadforge_build": 0.0,
2481
+ "cadforge_total": -1.0,
2482
+ "has_fixture": true,
2483
+ "has_cadquery_import": true,
2484
+ "error_kind": "TypeError"
2485
+ },
2486
+ {
2487
+ "reward_call": 52,
2488
+ "index": 3,
2489
+ "score": -0.79,
2490
+ "task_id": "desk_lamp_arm_joint",
2491
+ "code_chars": 6164,
2492
+ "cadforge_build": 0.0,
2493
+ "cadforge_total": -1.0,
2494
+ "has_fixture": true,
2495
+ "has_cadquery_import": true,
2496
+ "error_kind": "NameError"
2497
+ },
2498
+ {
2499
+ "reward_call": 53,
2500
+ "index": 0,
2501
+ "score": 0.8105635620789096,
2502
+ "task_id": "shaft_torque_clamp_120nm",
2503
+ "code_chars": 3731,
2504
+ "cadforge_build": 1.0,
2505
+ "cadforge_total": 0.7771336024457759,
2506
+ "has_fixture": true,
2507
+ "has_cadquery_import": true,
2508
+ "error_kind": "build_ok"
2509
+ },
2510
+ {
2511
+ "reward_call": 53,
2512
+ "index": 1,
2513
+ "score": 0.7106959361884481,
2514
+ "task_id": "shaft_torque_clamp_120nm",
2515
+ "code_chars": 4878,
2516
+ "cadforge_build": 1.0,
2517
+ "cadforge_total": 0.6596422778687624,
2518
+ "has_fixture": true,
2519
+ "has_cadquery_import": true,
2520
+ "error_kind": "build_ok"
2521
+ },
2522
+ {
2523
+ "reward_call": 53,
2524
+ "index": 2,
2525
+ "score": 0.6823218111628114,
2526
+ "task_id": "robot_gripper_two_finger",
2527
+ "code_chars": 5130,
2528
+ "cadforge_build": 1.0,
2529
+ "cadforge_total": 0.6262609543091899,
2530
+ "has_fixture": true,
2531
+ "has_cadquery_import": true,
2532
+ "error_kind": "build_ok"
2533
+ },
2534
+ {
2535
+ "reward_call": 53,
2536
+ "index": 3,
2537
+ "score": -0.645,
2538
+ "task_id": "robot_gripper_two_finger",
2539
+ "code_chars": 4339,
2540
+ "cadforge_build": 0.0,
2541
+ "cadforge_total": -1.0,
2542
+ "has_fixture": true,
2543
+ "has_cadquery_import": true,
2544
+ "error_kind": "NameError"
2545
+ },
2546
+ {
2547
+ "reward_call": 54,
2548
+ "index": 0,
2549
+ "score": -0.84,
2550
+ "task_id": "folding_step_stool",
2551
+ "code_chars": 5807,
2552
+ "cadforge_build": 0.0,
2553
+ "cadforge_total": -1.0,
2554
+ "has_fixture": false,
2555
+ "has_cadquery_import": true,
2556
+ "error_kind": "SyntaxError"
2557
+ },
2558
+ {
2559
+ "reward_call": 54,
2560
+ "index": 1,
2561
+ "score": -0.84,
2562
+ "task_id": "folding_step_stool",
2563
+ "code_chars": 6139,
2564
+ "cadforge_build": 0.0,
2565
+ "cadforge_total": -1.0,
2566
+ "has_fixture": false,
2567
+ "has_cadquery_import": true,
2568
+ "error_kind": "SyntaxError"
2569
+ },
2570
+ {
2571
+ "reward_call": 54,
2572
+ "index": 2,
2573
+ "score": 0.7464098823051949,
2574
+ "task_id": "pipe_clamp_saddle",
2575
+ "code_chars": 5033,
2576
+ "cadforge_build": 1.0,
2577
+ "cadforge_total": 0.7016586850649351,
2578
+ "has_fixture": true,
2579
+ "has_cadquery_import": true,
2580
+ "error_kind": "build_ok"
2581
+ },
2582
+ {
2583
+ "reward_call": 54,
2584
+ "index": 3,
2585
+ "score": -0.565,
2586
+ "task_id": "pipe_clamp_saddle",
2587
+ "code_chars": 4372,
2588
+ "cadforge_build": 0.0,
2589
+ "cadforge_total": -1.0,
2590
+ "has_fixture": true,
2591
+ "has_cadquery_import": true,
2592
+ "error_kind": "TypeError"
2593
+ },
2594
+ {
2595
+ "reward_call": 55,
2596
+ "index": 0,
2597
+ "score": 0.6493090788741098,
2598
+ "task_id": "wall_j_hook_120n",
2599
+ "code_chars": 3841,
2600
+ "cadforge_build": 1.0,
2601
+ "cadforge_total": 0.5874224457342468,
2602
+ "has_fixture": true,
2603
+ "has_cadquery_import": true,
2604
+ "error_kind": "build_ok"
2605
+ },
2606
+ {
2607
+ "reward_call": 55,
2608
+ "index": 1,
2609
+ "score": -0.565,
2610
+ "task_id": "wall_j_hook_120n",
2611
+ "code_chars": 3368,
2612
+ "cadforge_build": 0.0,
2613
+ "cadforge_total": -1.0,
2614
+ "has_fixture": true,
2615
+ "has_cadquery_import": true,
2616
+ "error_kind": "ValueError"
2617
+ },
2618
+ {
2619
+ "reward_call": 55,
2620
+ "index": 2,
2621
+ "score": -0.84,
2622
+ "task_id": "gearbox_housing_split",
2623
+ "code_chars": 5987,
2624
+ "cadforge_build": 0.0,
2625
+ "cadforge_total": -1.0,
2626
+ "has_fixture": false,
2627
+ "has_cadquery_import": true,
2628
+ "error_kind": "SyntaxError"
2629
+ },
2630
+ {
2631
+ "reward_call": 55,
2632
+ "index": 3,
2633
+ "score": -0.8699999999999999,
2634
+ "task_id": "gearbox_housing_split",
2635
+ "code_chars": 5467,
2636
+ "cadforge_build": 0.0,
2637
+ "cadforge_total": -1.0,
2638
+ "has_fixture": false,
2639
+ "has_cadquery_import": true,
2640
+ "error_kind": "SyntaxError"
2641
+ },
2642
+ {
2643
+ "reward_call": 56,
2644
+ "index": 0,
2645
+ "score": -0.84,
2646
+ "task_id": "lightweight_truss_support_250n",
2647
+ "code_chars": 4523,
2648
+ "cadforge_build": 0.0,
2649
+ "cadforge_total": -1.0,
2650
+ "has_fixture": false,
2651
+ "has_cadquery_import": true,
2652
+ "error_kind": "SyntaxError"
2653
+ },
2654
+ {
2655
+ "reward_call": 56,
2656
+ "index": 1,
2657
+ "score": -0.645,
2658
+ "task_id": "lightweight_truss_support_250n",
2659
+ "code_chars": 4927,
2660
+ "cadforge_build": 0.0,
2661
+ "cadforge_total": -1.0,
2662
+ "has_fixture": true,
2663
+ "has_cadquery_import": true,
2664
+ "error_kind": "AttributeError"
2665
+ },
2666
+ {
2667
+ "reward_call": 56,
2668
+ "index": 2,
2669
+ "score": 0.7611631331328613,
2670
+ "task_id": "laptop_riser_vented",
2671
+ "code_chars": 3601,
2672
+ "cadforge_build": 1.0,
2673
+ "cadforge_total": 0.7190154507445428,
2674
+ "has_fixture": true,
2675
+ "has_cadquery_import": true,
2676
+ "error_kind": "build_ok"
2677
+ },
2678
+ {
2679
+ "reward_call": 56,
2680
+ "index": 3,
2681
+ "score": 0.7619837729036159,
2682
+ "task_id": "laptop_riser_vented",
2683
+ "code_chars": 3894,
2684
+ "cadforge_build": 1.0,
2685
+ "cadforge_total": 0.7199809092983716,
2686
+ "has_fixture": true,
2687
+ "has_cadquery_import": true,
2688
+ "error_kind": "build_ok"
2689
+ },
2690
+ {
2691
+ "reward_call": 57,
2692
+ "index": 0,
2693
+ "score": -0.8849999999999999,
2694
+ "task_id": "bench_vise_simplified",
2695
+ "code_chars": 5492,
2696
+ "cadforge_build": 0.0,
2697
+ "cadforge_total": -1.0,
2698
+ "has_fixture": false,
2699
+ "has_cadquery_import": true,
2700
+ "error_kind": "SyntaxError"
2701
+ },
2702
+ {
2703
+ "reward_call": 57,
2704
+ "index": 1,
2705
+ "score": -0.7849999999999999,
2706
+ "task_id": "bench_vise_simplified",
2707
+ "code_chars": 5359,
2708
+ "cadforge_build": 0.0,
2709
+ "cadforge_total": -1.0,
2710
+ "has_fixture": false,
2711
+ "has_cadquery_import": true,
2712
+ "error_kind": "ValueError"
2713
+ },
2714
+ {
2715
+ "reward_call": 57,
2716
+ "index": 2,
2717
+ "score": 0.7362512650469238,
2718
+ "task_id": "micro_drone_frame",
2719
+ "code_chars": 5453,
2720
+ "cadforge_build": 1.0,
2721
+ "cadforge_total": 0.6897073706434398,
2722
+ "has_fixture": true,
2723
+ "has_cadquery_import": true,
2724
+ "error_kind": "build_ok"
2725
+ },
2726
+ {
2727
+ "reward_call": 57,
2728
+ "index": 3,
2729
+ "score": -0.84,
2730
+ "task_id": "micro_drone_frame",
2731
+ "code_chars": 5941,
2732
+ "cadforge_build": 0.0,
2733
+ "cadforge_total": -1.0,
2734
+ "has_fixture": false,
2735
+ "has_cadquery_import": true,
2736
+ "error_kind": "SyntaxError"
2737
+ },
2738
+ {
2739
+ "reward_call": 58,
2740
+ "index": 0,
2741
+ "score": -0.84,
2742
+ "task_id": "ergonomic_curvy_chair_1000n",
2743
+ "code_chars": 4919,
2744
+ "cadforge_build": 0.0,
2745
+ "cadforge_total": -1.0,
2746
+ "has_fixture": false,
2747
+ "has_cadquery_import": true,
2748
+ "error_kind": "SyntaxError"
2749
+ },
2750
+ {
2751
+ "reward_call": 58,
2752
+ "index": 1,
2753
+ "score": -0.74,
2754
+ "task_id": "ergonomic_curvy_chair_1000n",
2755
+ "code_chars": 5406,
2756
+ "cadforge_build": 0.0,
2757
+ "cadforge_total": -1.0,
2758
+ "has_fixture": false,
2759
+ "has_cadquery_import": true,
2760
+ "error_kind": "ValueError"
2761
+ },
2762
+ {
2763
+ "reward_call": 58,
2764
+ "index": 2,
2765
+ "score": 0.889293602353931,
2766
+ "task_id": "phone_stand_adjustable_hinge",
2767
+ "code_chars": 4588,
2768
+ "cadforge_build": 1.0,
2769
+ "cadforge_total": 0.8697571792399188,
2770
+ "has_fixture": true,
2771
+ "has_cadquery_import": true,
2772
+ "error_kind": "build_ok"
2773
+ },
2774
+ {
2775
+ "reward_call": 58,
2776
+ "index": 3,
2777
+ "score": 0.7180833333333333,
2778
+ "task_id": "phone_stand_adjustable_hinge",
2779
+ "code_chars": 3710,
2780
+ "cadforge_build": 1.0,
2781
+ "cadforge_total": 0.6683333333333333,
2782
+ "has_fixture": true,
2783
+ "has_cadquery_import": true,
2784
+ "error_kind": "build_ok"
2785
+ },
2786
+ {
2787
+ "reward_call": 59,
2788
+ "index": 0,
2789
+ "score": -0.6649999999999999,
2790
+ "task_id": "table_six_leg_500n",
2791
+ "code_chars": 3985,
2792
+ "cadforge_build": 0.0,
2793
+ "cadforge_total": -1.0,
2794
+ "has_fixture": true,
2795
+ "has_cadquery_import": true,
2796
+ "error_kind": "SyntaxError"
2797
+ },
2798
+ {
2799
+ "reward_call": 59,
2800
+ "index": 1,
2801
+ "score": 0.7751041666666667,
2802
+ "task_id": "table_six_leg_500n",
2803
+ "code_chars": 4474,
2804
+ "cadforge_build": 1.0,
2805
+ "cadforge_total": 0.7354166666666667,
2806
+ "has_fixture": true,
2807
+ "has_cadquery_import": true,
2808
+ "error_kind": "build_ok"
2809
+ },
2810
+ {
2811
+ "reward_call": 59,
2812
+ "index": 2,
2813
+ "score": -0.565,
2814
+ "task_id": "bike_accessory_mount_120n",
2815
+ "code_chars": 3449,
2816
+ "cadforge_build": 0.0,
2817
+ "cadforge_total": -1.0,
2818
+ "has_fixture": true,
2819
+ "has_cadquery_import": true,
2820
+ "error_kind": "unknown"
2821
+ },
2822
+ {
2823
+ "reward_call": 59,
2824
+ "index": 3,
2825
+ "score": -0.6649999999999999,
2826
+ "task_id": "bike_accessory_mount_120n",
2827
+ "code_chars": 5347,
2828
+ "cadforge_build": 0.0,
2829
+ "cadforge_total": -1.0,
2830
+ "has_fixture": true,
2831
+ "has_cadquery_import": true,
2832
+ "error_kind": "SyntaxError"
2833
+ },
2834
+ {
2835
+ "reward_call": 60,
2836
+ "index": 0,
2837
+ "score": 0.7578764590434719,
2838
+ "task_id": "bottle_cage_bike",
2839
+ "code_chars": 3193,
2840
+ "cadforge_build": 1.0,
2841
+ "cadforge_total": 0.715148775345261,
2842
+ "has_fixture": true,
2843
+ "has_cadquery_import": true,
2844
+ "error_kind": "build_ok"
2845
+ },
2846
+ {
2847
+ "reward_call": 60,
2848
+ "index": 1,
2849
+ "score": -0.615,
2850
+ "task_id": "bottle_cage_bike",
2851
+ "code_chars": 4094,
2852
+ "cadforge_build": 0.0,
2853
+ "cadforge_total": -1.0,
2854
+ "has_fixture": true,
2855
+ "has_cadquery_import": true,
2856
+ "error_kind": "AttributeError"
2857
+ },
2858
+ {
2859
+ "reward_call": 60,
2860
+ "index": 2,
2861
+ "score": -0.6649999999999999,
2862
+ "task_id": "cable_clip_snap",
2863
+ "code_chars": 5720,
2864
+ "cadforge_build": 0.0,
2865
+ "cadforge_total": -1.0,
2866
+ "has_fixture": true,
2867
+ "has_cadquery_import": true,
2868
+ "error_kind": "SyntaxError"
2869
+ },
2870
+ {
2871
+ "reward_call": 60,
2872
+ "index": 3,
2873
+ "score": 0.7115461458630677,
2874
+ "task_id": "cable_clip_snap",
2875
+ "code_chars": 4541,
2876
+ "cadforge_build": 1.0,
2877
+ "cadforge_total": 0.6606425245447856,
2878
+ "has_fixture": true,
2879
+ "has_cadquery_import": true,
2880
+ "error_kind": "build_ok"
2881
+ },
2882
+ {
2883
+ "reward_call": 61,
2884
+ "index": 0,
2885
+ "score": -0.61,
2886
+ "task_id": "laptop_riser_vented",
2887
+ "code_chars": 5017,
2888
+ "cadforge_build": 0.0,
2889
+ "cadforge_total": -1.0,
2890
+ "has_fixture": true,
2891
+ "has_cadquery_import": true,
2892
+ "error_kind": "TypeError"
2893
+ },
2894
+ {
2895
+ "reward_call": 61,
2896
+ "index": 1,
2897
+ "score": 0.8574772795845026,
2898
+ "task_id": "laptop_riser_vented",
2899
+ "code_chars": 4781,
2900
+ "cadforge_build": 1.0,
2901
+ "cadforge_total": 0.8323262112758855,
2902
+ "has_fixture": true,
2903
+ "has_cadquery_import": true,
2904
+ "error_kind": "build_ok"
2905
+ },
2906
+ {
2907
+ "reward_call": 61,
2908
+ "index": 2,
2909
+ "score": -0.615,
2910
+ "task_id": "wall_j_hook_120n",
2911
+ "code_chars": 4160,
2912
+ "cadforge_build": 0.0,
2913
+ "cadforge_total": -1.0,
2914
+ "has_fixture": true,
2915
+ "has_cadquery_import": true,
2916
+ "error_kind": "NameError"
2917
+ },
2918
+ {
2919
+ "reward_call": 61,
2920
+ "index": 3,
2921
+ "score": -0.615,
2922
+ "task_id": "wall_j_hook_120n",
2923
+ "code_chars": 2626,
2924
+ "cadforge_build": 0.0,
2925
+ "cadforge_total": -1.0,
2926
+ "has_fixture": true,
2927
+ "has_cadquery_import": true,
2928
+ "error_kind": "AttributeError"
2929
+ },
2930
+ {
2931
+ "reward_call": 62,
2932
+ "index": 0,
2933
+ "score": -0.565,
2934
+ "task_id": "desk_lamp_arm_joint",
2935
+ "code_chars": 5457,
2936
+ "cadforge_build": 0.0,
2937
+ "cadforge_total": -1.0,
2938
+ "has_fixture": true,
2939
+ "has_cadquery_import": true,
2940
+ "error_kind": "TypeError"
2941
+ },
2942
+ {
2943
+ "reward_call": 62,
2944
+ "index": 1,
2945
+ "score": -0.84,
2946
+ "task_id": "desk_lamp_arm_joint",
2947
+ "code_chars": 5326,
2948
+ "cadforge_build": 0.0,
2949
+ "cadforge_total": -1.0,
2950
+ "has_fixture": false,
2951
+ "has_cadquery_import": true,
2952
+ "error_kind": "SyntaxError"
2953
+ },
2954
+ {
2955
+ "reward_call": 62,
2956
+ "index": 2,
2957
+ "score": -0.6649999999999999,
2958
+ "task_id": "cable_clip_snap",
2959
+ "code_chars": 5669,
2960
+ "cadforge_build": 0.0,
2961
+ "cadforge_total": -1.0,
2962
+ "has_fixture": true,
2963
+ "has_cadquery_import": true,
2964
+ "error_kind": "SyntaxError"
2965
+ },
2966
+ {
2967
+ "reward_call": 62,
2968
+ "index": 3,
2969
+ "score": -0.66,
2970
+ "task_id": "cable_clip_snap",
2971
+ "code_chars": 4840,
2972
+ "cadforge_build": 0.0,
2973
+ "cadforge_total": -1.0,
2974
+ "has_fixture": true,
2975
+ "has_cadquery_import": true,
2976
+ "error_kind": "NameError"
2977
+ },
2978
+ {
2979
+ "reward_call": 63,
2980
+ "index": 0,
2981
+ "score": -0.6649999999999999,
2982
+ "task_id": "lightweight_truss_support_250n",
2983
+ "code_chars": 5609,
2984
+ "cadforge_build": 0.0,
2985
+ "cadforge_total": -1.0,
2986
+ "has_fixture": true,
2987
+ "has_cadquery_import": true,
2988
+ "error_kind": "SyntaxError"
2989
+ },
2990
+ {
2991
+ "reward_call": 63,
2992
+ "index": 1,
2993
+ "score": 0.7126822916666666,
2994
+ "task_id": "lightweight_truss_support_250n",
2995
+ "code_chars": 3365,
2996
+ "cadforge_build": 1.0,
2997
+ "cadforge_total": 0.6619791666666667,
2998
+ "has_fixture": true,
2999
+ "has_cadquery_import": true,
3000
+ "error_kind": "build_ok"
3001
+ },
3002
+ {
3003
+ "reward_call": 63,
3004
+ "index": 2,
3005
+ "score": -0.84,
3006
+ "task_id": "gearbox_housing_split",
3007
+ "code_chars": 6067,
3008
+ "cadforge_build": 0.0,
3009
+ "cadforge_total": -1.0,
3010
+ "has_fixture": false,
3011
+ "has_cadquery_import": true,
3012
+ "error_kind": "SyntaxError"
3013
+ },
3014
+ {
3015
+ "reward_call": 63,
3016
+ "index": 3,
3017
+ "score": -0.84,
3018
+ "task_id": "gearbox_housing_split",
3019
+ "code_chars": 5620,
3020
+ "cadforge_build": 0.0,
3021
+ "cadforge_total": -1.0,
3022
+ "has_fixture": false,
3023
+ "has_cadquery_import": true,
3024
+ "error_kind": "SyntaxError"
3025
+ },
3026
+ {
3027
+ "reward_call": 64,
3028
+ "index": 0,
3029
+ "score": -0.615,
3030
+ "task_id": "bottle_cage_bike",
3031
+ "code_chars": 4779,
3032
+ "cadforge_build": 0.0,
3033
+ "cadforge_total": -1.0,
3034
+ "has_fixture": true,
3035
+ "has_cadquery_import": true,
3036
+ "error_kind": "NameError"
3037
+ },
3038
+ {
3039
+ "reward_call": 64,
3040
+ "index": 1,
3041
+ "score": -0.615,
3042
+ "task_id": "bottle_cage_bike",
3043
+ "code_chars": 6025,
3044
+ "cadforge_build": 0.0,
3045
+ "cadforge_total": -1.0,
3046
+ "has_fixture": true,
3047
+ "has_cadquery_import": true,
3048
+ "error_kind": "NameError"
3049
+ },
3050
+ {
3051
+ "reward_call": 64,
3052
+ "index": 2,
3053
+ "score": -0.74,
3054
+ "task_id": "micro_drone_frame",
3055
+ "code_chars": 5747,
3056
+ "cadforge_build": 0.0,
3057
+ "cadforge_total": -1.0,
3058
+ "has_fixture": false,
3059
+ "has_cadquery_import": true,
3060
+ "error_kind": "ValueError"
3061
+ },
3062
+ {
3063
+ "reward_call": 64,
3064
+ "index": 3,
3065
+ "score": -0.615,
3066
+ "task_id": "micro_drone_frame",
3067
+ "code_chars": 5582,
3068
+ "cadforge_build": 0.0,
3069
+ "cadforge_total": -1.0,
3070
+ "has_fixture": true,
3071
+ "has_cadquery_import": true,
3072
+ "error_kind": "AttributeError"
3073
+ },
3074
+ {
3075
+ "reward_call": 65,
3076
+ "index": 0,
3077
+ "score": -0.74,
3078
+ "task_id": "phone_stand_adjustable_hinge",
3079
+ "code_chars": 5572,
3080
+ "cadforge_build": 0.0,
3081
+ "cadforge_total": -1.0,
3082
+ "has_fixture": false,
3083
+ "has_cadquery_import": true,
3084
+ "error_kind": "ValueError"
3085
+ },
3086
+ {
3087
+ "reward_call": 65,
3088
+ "index": 1,
3089
+ "score": 0.9449412121370865,
3090
+ "task_id": "phone_stand_adjustable_hinge",
3091
+ "code_chars": 3965,
3092
+ "cadforge_build": 1.0,
3093
+ "cadforge_total": 0.9352249554553959,
3094
+ "has_fixture": true,
3095
+ "has_cadquery_import": true,
3096
+ "error_kind": "build_ok"
3097
+ },
3098
+ {
3099
+ "reward_call": 65,
3100
+ "index": 2,
3101
+ "score": -0.61,
3102
+ "task_id": "robot_gripper_two_finger",
3103
+ "code_chars": 4726,
3104
+ "cadforge_build": 0.0,
3105
+ "cadforge_total": -1.0,
3106
+ "has_fixture": true,
3107
+ "has_cadquery_import": true,
3108
+ "error_kind": "unknown"
3109
+ },
3110
+ {
3111
+ "reward_call": 65,
3112
+ "index": 3,
3113
+ "score": 0.72758080696295,
3114
+ "task_id": "robot_gripper_two_finger",
3115
+ "code_chars": 3254,
3116
+ "cadforge_build": 1.0,
3117
+ "cadforge_total": 0.6795068317211177,
3118
+ "has_fixture": true,
3119
+ "has_cadquery_import": true,
3120
+ "error_kind": "build_ok"
3121
+ },
3122
+ {
3123
+ "reward_call": 66,
3124
+ "index": 0,
3125
+ "score": 0.6171452781088472,
3126
+ "task_id": "bench_vise_simplified",
3127
+ "code_chars": 5564,
3128
+ "cadforge_build": 1.0,
3129
+ "cadforge_total": 0.5495826801280556,
3130
+ "has_fixture": true,
3131
+ "has_cadquery_import": true,
3132
+ "error_kind": "build_ok"
3133
+ },
3134
+ {
3135
+ "reward_call": 66,
3136
+ "index": 1,
3137
+ "score": -0.84,
3138
+ "task_id": "bench_vise_simplified",
3139
+ "code_chars": 5617,
3140
+ "cadforge_build": 0.0,
3141
+ "cadforge_total": -1.0,
3142
+ "has_fixture": false,
3143
+ "has_cadquery_import": true,
3144
+ "error_kind": "SyntaxError"
3145
+ },
3146
+ {
3147
+ "reward_call": 66,
3148
+ "index": 2,
3149
+ "score": 0.7287083333333334,
3150
+ "task_id": "pipe_clamp_saddle",
3151
+ "code_chars": 4561,
3152
+ "cadforge_build": 1.0,
3153
+ "cadforge_total": 0.6808333333333334,
3154
+ "has_fixture": true,
3155
+ "has_cadquery_import": true,
3156
+ "error_kind": "build_ok"
3157
+ },
3158
+ {
3159
+ "reward_call": 66,
3160
+ "index": 3,
3161
+ "score": -0.6900000000000001,
3162
+ "task_id": "pipe_clamp_saddle",
3163
+ "code_chars": 3756,
3164
+ "cadforge_build": 0.0,
3165
+ "cadforge_total": -1.0,
3166
+ "has_fixture": true,
3167
+ "has_cadquery_import": true,
3168
+ "error_kind": "NameError"
3169
+ },
3170
+ {
3171
+ "reward_call": 67,
3172
+ "index": 0,
3173
+ "score": -0.595,
3174
+ "task_id": "drawer_handle_two_bolt",
3175
+ "code_chars": 3019,
3176
+ "cadforge_build": 0.0,
3177
+ "cadforge_total": -1.0,
3178
+ "has_fixture": true,
3179
+ "has_cadquery_import": true,
3180
+ "error_kind": "unknown"
3181
+ },
3182
+ {
3183
+ "reward_call": 67,
3184
+ "index": 1,
3185
+ "score": 0.7683493216086832,
3186
+ "task_id": "drawer_handle_two_bolt",
3187
+ "code_chars": 2128,
3188
+ "cadforge_build": 1.0,
3189
+ "cadforge_total": 0.7274697901278626,
3190
+ "has_fixture": true,
3191
+ "has_cadquery_import": true,
3192
+ "error_kind": "build_ok"
3193
+ },
3194
+ {
3195
+ "reward_call": 67,
3196
+ "index": 2,
3197
+ "score": -0.565,
3198
+ "task_id": "hinge_leaf_barrel",
3199
+ "code_chars": 4949,
3200
+ "cadforge_build": 0.0,
3201
+ "cadforge_total": -1.0,
3202
+ "has_fixture": true,
3203
+ "has_cadquery_import": true,
3204
+ "error_kind": "unknown"
3205
+ },
3206
+ {
3207
+ "reward_call": 67,
3208
+ "index": 3,
3209
+ "score": -0.595,
3210
+ "task_id": "hinge_leaf_barrel",
3211
+ "code_chars": 5208,
3212
+ "cadforge_build": 0.0,
3213
+ "cadforge_total": -1.0,
3214
+ "has_fixture": true,
3215
+ "has_cadquery_import": true,
3216
+ "error_kind": "TypeError"
3217
+ },
3218
+ {
3219
+ "reward_call": 68,
3220
+ "index": 0,
3221
+ "score": 0.7240374070978418,
3222
+ "task_id": "shelf_bracket_triangular_200n",
3223
+ "code_chars": 2343,
3224
+ "cadforge_build": 1.0,
3225
+ "cadforge_total": 0.6753381259974609,
3226
+ "has_fixture": true,
3227
+ "has_cadquery_import": true,
3228
+ "error_kind": "build_ok"
3229
+ },
3230
+ {
3231
+ "reward_call": 68,
3232
+ "index": 1,
3233
+ "score": 0.81054143511043,
3234
+ "task_id": "shelf_bracket_triangular_200n",
3235
+ "code_chars": 3230,
3236
+ "cadforge_build": 1.0,
3237
+ "cadforge_total": 0.777107570718153,
3238
+ "has_fixture": true,
3239
+ "has_cadquery_import": true,
3240
+ "error_kind": "build_ok"
3241
+ },
3242
+ {
3243
+ "reward_call": 68,
3244
+ "index": 2,
3245
+ "score": -0.6649999999999999,
3246
+ "task_id": "shaft_torque_clamp_120nm",
3247
+ "code_chars": 5108,
3248
+ "cadforge_build": 0.0,
3249
+ "cadforge_total": -1.0,
3250
+ "has_fixture": true,
3251
+ "has_cadquery_import": true,
3252
+ "error_kind": "SyntaxError"
3253
+ },
3254
+ {
3255
+ "reward_call": 68,
3256
+ "index": 3,
3257
+ "score": -0.565,
3258
+ "task_id": "shaft_torque_clamp_120nm",
3259
+ "code_chars": 4955,
3260
+ "cadforge_build": 0.0,
3261
+ "cadforge_total": -1.0,
3262
+ "has_fixture": true,
3263
+ "has_cadquery_import": true,
3264
+ "error_kind": "unknown"
3265
+ },
3266
+ {
3267
+ "reward_call": 69,
3268
+ "index": 0,
3269
+ "score": -0.565,
3270
+ "task_id": "bike_accessory_mount_120n",
3271
+ "code_chars": 4445,
3272
+ "cadforge_build": 0.0,
3273
+ "cadforge_total": -1.0,
3274
+ "has_fixture": true,
3275
+ "has_cadquery_import": true,
3276
+ "error_kind": "BRep_API"
3277
+ },
3278
+ {
3279
+ "reward_call": 69,
3280
+ "index": 1,
3281
+ "score": 0.76625,
3282
+ "task_id": "bike_accessory_mount_120n",
3283
+ "code_chars": 3653,
3284
+ "cadforge_build": 1.0,
3285
+ "cadforge_total": 0.725,
3286
+ "has_fixture": true,
3287
+ "has_cadquery_import": true,
3288
+ "error_kind": "build_ok"
3289
+ },
3290
+ {
3291
+ "reward_call": 69,
3292
+ "index": 2,
3293
+ "score": -0.875,
3294
+ "task_id": "folding_step_stool",
3295
+ "code_chars": 5125,
3296
+ "cadforge_build": 0.0,
3297
+ "cadforge_total": -1.0,
3298
+ "has_fixture": false,
3299
+ "has_cadquery_import": false,
3300
+ "error_kind": "ValueError"
3301
+ },
3302
+ {
3303
+ "reward_call": 69,
3304
+ "index": 3,
3305
+ "score": -0.84,
3306
+ "task_id": "folding_step_stool",
3307
+ "code_chars": 6132,
3308
+ "cadforge_build": 0.0,
3309
+ "cadforge_total": -1.0,
3310
+ "has_fixture": false,
3311
+ "has_cadquery_import": true,
3312
+ "error_kind": "SyntaxError"
3313
+ },
3314
+ {
3315
+ "reward_call": 70,
3316
+ "index": 0,
3317
+ "score": 0.7472510593220338,
3318
+ "task_id": "table_six_leg_500n",
3319
+ "code_chars": 4603,
3320
+ "cadforge_build": 1.0,
3321
+ "cadforge_total": 0.7026483050847457,
3322
+ "has_fixture": true,
3323
+ "has_cadquery_import": true,
3324
+ "error_kind": "build_ok"
3325
+ },
3326
+ {
3327
+ "reward_call": 70,
3328
+ "index": 1,
3329
+ "score": 0.8660160585254221,
3330
+ "task_id": "table_six_leg_500n",
3331
+ "code_chars": 3278,
3332
+ "cadforge_build": 1.0,
3333
+ "cadforge_total": 0.8423718335593201,
3334
+ "has_fixture": true,
3335
+ "has_cadquery_import": true,
3336
+ "error_kind": "build_ok"
3337
+ },
3338
+ {
3339
+ "reward_call": 70,
3340
+ "index": 2,
3341
+ "score": -0.79,
3342
+ "task_id": "ergonomic_curvy_chair_1000n",
3343
+ "code_chars": 4733,
3344
+ "cadforge_build": 0.0,
3345
+ "cadforge_total": -1.0,
3346
+ "has_fixture": false,
3347
+ "has_cadquery_import": true,
3348
+ "error_kind": "NameError"
3349
+ },
3350
+ {
3351
+ "reward_call": 70,
3352
+ "index": 3,
3353
+ "score": -0.84,
3354
+ "task_id": "ergonomic_curvy_chair_1000n",
3355
+ "code_chars": 4873,
3356
+ "cadforge_build": 0.0,
3357
+ "cadforge_total": -1.0,
3358
+ "has_fixture": false,
3359
+ "has_cadquery_import": true,
3360
+ "error_kind": "SyntaxError"
3361
+ },
3362
+ {
3363
+ "reward_call": 71,
3364
+ "index": 0,
3365
+ "score": -0.565,
3366
+ "task_id": "bench_vise_simplified",
3367
+ "code_chars": 4576,
3368
+ "cadforge_build": 0.0,
3369
+ "cadforge_total": -1.0,
3370
+ "has_fixture": true,
3371
+ "has_cadquery_import": true,
3372
+ "error_kind": "ValueError"
3373
+ },
3374
+ {
3375
+ "reward_call": 71,
3376
+ "index": 1,
3377
+ "score": 0.7258088453554393,
3378
+ "task_id": "bench_vise_simplified",
3379
+ "code_chars": 4090,
3380
+ "cadforge_build": 1.0,
3381
+ "cadforge_total": 0.6774221710063992,
3382
+ "has_fixture": true,
3383
+ "has_cadquery_import": true,
3384
+ "error_kind": "build_ok"
3385
+ },
3386
+ {
3387
+ "reward_call": 71,
3388
+ "index": 2,
3389
+ "score": -0.6649999999999999,
3390
+ "task_id": "drawer_handle_two_bolt",
3391
+ "code_chars": 2945,
3392
+ "cadforge_build": 0.0,
3393
+ "cadforge_total": -1.0,
3394
+ "has_fixture": true,
3395
+ "has_cadquery_import": true,
3396
+ "error_kind": "SyntaxError"
3397
+ },
3398
+ {
3399
+ "reward_call": 71,
3400
+ "index": 3,
3401
+ "score": -0.645,
3402
+ "task_id": "drawer_handle_two_bolt",
3403
+ "code_chars": 2655,
3404
+ "cadforge_build": 0.0,
3405
+ "cadforge_total": -1.0,
3406
+ "has_fixture": true,
3407
+ "has_cadquery_import": true,
3408
+ "error_kind": "NameError"
3409
+ },
3410
+ {
3411
+ "reward_call": 72,
3412
+ "index": 0,
3413
+ "score": -0.84,
3414
+ "task_id": "ergonomic_curvy_chair_1000n",
3415
+ "code_chars": 5184,
3416
+ "cadforge_build": 0.0,
3417
+ "cadforge_total": -1.0,
3418
+ "has_fixture": false,
3419
+ "has_cadquery_import": true,
3420
+ "error_kind": "SyntaxError"
3421
+ },
3422
+ {
3423
+ "reward_call": 72,
3424
+ "index": 1,
3425
+ "score": -0.84,
3426
+ "task_id": "ergonomic_curvy_chair_1000n",
3427
+ "code_chars": 5287,
3428
+ "cadforge_build": 0.0,
3429
+ "cadforge_total": -1.0,
3430
+ "has_fixture": false,
3431
+ "has_cadquery_import": true,
3432
+ "error_kind": "SyntaxError"
3433
+ },
3434
+ {
3435
+ "reward_call": 72,
3436
+ "index": 2,
3437
+ "score": -0.615,
3438
+ "task_id": "laptop_riser_vented",
3439
+ "code_chars": 4588,
3440
+ "cadforge_build": 0.0,
3441
+ "cadforge_total": -1.0,
3442
+ "has_fixture": true,
3443
+ "has_cadquery_import": true,
3444
+ "error_kind": "AttributeError"
3445
+ },
3446
+ {
3447
+ "reward_call": 72,
3448
+ "index": 3,
3449
+ "score": 0.8629728660675577,
3450
+ "task_id": "laptop_riser_vented",
3451
+ "code_chars": 3536,
3452
+ "cadforge_build": 1.0,
3453
+ "cadforge_total": 0.8387916071383033,
3454
+ "has_fixture": true,
3455
+ "has_cadquery_import": true,
3456
+ "error_kind": "build_ok"
3457
+ },
3458
+ {
3459
+ "reward_call": 73,
3460
+ "index": 0,
3461
+ "score": -0.6649999999999999,
3462
+ "task_id": "micro_drone_frame",
3463
+ "code_chars": 5797,
3464
+ "cadforge_build": 0.0,
3465
+ "cadforge_total": -1.0,
3466
+ "has_fixture": true,
3467
+ "has_cadquery_import": true,
3468
+ "error_kind": "SyntaxError"
3469
+ },
3470
+ {
3471
+ "reward_call": 73,
3472
+ "index": 1,
3473
+ "score": -0.84,
3474
+ "task_id": "micro_drone_frame",
3475
+ "code_chars": 5772,
3476
+ "cadforge_build": 0.0,
3477
+ "cadforge_total": -1.0,
3478
+ "has_fixture": false,
3479
+ "has_cadquery_import": true,
3480
+ "error_kind": "SyntaxError"
3481
+ },
3482
+ {
3483
+ "reward_call": 73,
3484
+ "index": 2,
3485
+ "score": 0.8132510351803028,
3486
+ "task_id": "bike_accessory_mount_120n",
3487
+ "code_chars": 5776,
3488
+ "cadforge_build": 1.0,
3489
+ "cadforge_total": 0.7802953355062386,
3490
+ "has_fixture": true,
3491
+ "has_cadquery_import": true,
3492
+ "error_kind": "build_ok"
3493
+ },
3494
+ {
3495
+ "reward_call": 73,
3496
+ "index": 3,
3497
+ "score": -0.565,
3498
+ "task_id": "bike_accessory_mount_120n",
3499
+ "code_chars": 5097,
3500
+ "cadforge_build": 0.0,
3501
+ "cadforge_total": -1.0,
3502
+ "has_fixture": true,
3503
+ "has_cadquery_import": true,
3504
+ "error_kind": "TypeError"
3505
+ },
3506
+ {
3507
+ "reward_call": 74,
3508
+ "index": 0,
3509
+ "score": 0.7389633309206752,
3510
+ "task_id": "desk_lamp_arm_joint",
3511
+ "code_chars": 5575,
3512
+ "cadforge_build": 1.0,
3513
+ "cadforge_total": 0.692898036377265,
3514
+ "has_fixture": true,
3515
+ "has_cadquery_import": true,
3516
+ "error_kind": "build_ok"
3517
+ },
3518
+ {
3519
+ "reward_call": 74,
3520
+ "index": 1,
3521
+ "score": -0.84,
3522
+ "task_id": "desk_lamp_arm_joint",
3523
+ "code_chars": 5890,
3524
+ "cadforge_build": 0.0,
3525
+ "cadforge_total": -1.0,
3526
+ "has_fixture": false,
3527
+ "has_cadquery_import": true,
3528
+ "error_kind": "SyntaxError"
3529
+ },
3530
+ {
3531
+ "reward_call": 74,
3532
+ "index": 2,
3533
+ "score": -0.595,
3534
+ "task_id": "hinge_leaf_barrel",
3535
+ "code_chars": 3936,
3536
+ "cadforge_build": 0.0,
3537
+ "cadforge_total": -1.0,
3538
+ "has_fixture": true,
3539
+ "has_cadquery_import": true,
3540
+ "error_kind": "TypeError"
3541
+ },
3542
+ {
3543
+ "reward_call": 74,
3544
+ "index": 3,
3545
+ "score": -0.595,
3546
+ "task_id": "hinge_leaf_barrel",
3547
+ "code_chars": 3966,
3548
+ "cadforge_build": 0.0,
3549
+ "cadforge_total": -1.0,
3550
+ "has_fixture": true,
3551
+ "has_cadquery_import": true,
3552
+ "error_kind": "TypeError"
3553
+ },
3554
+ {
3555
+ "reward_call": 75,
3556
+ "index": 0,
3557
+ "score": -0.84,
3558
+ "task_id": "gearbox_housing_split",
3559
+ "code_chars": 6076,
3560
+ "cadforge_build": 0.0,
3561
+ "cadforge_total": -1.0,
3562
+ "has_fixture": false,
3563
+ "has_cadquery_import": true,
3564
+ "error_kind": "SyntaxError"
3565
+ },
3566
+ {
3567
+ "reward_call": 75,
3568
+ "index": 1,
3569
+ "score": 0.8263909278635598,
3570
+ "task_id": "gearbox_housing_split",
3571
+ "code_chars": 5652,
3572
+ "cadforge_build": 1.0,
3573
+ "cadforge_total": 0.7957540327806586,
3574
+ "has_fixture": true,
3575
+ "has_cadquery_import": true,
3576
+ "error_kind": "build_ok"
3577
+ },
3578
+ {
3579
+ "reward_call": 75,
3580
+ "index": 2,
3581
+ "score": -0.645,
3582
+ "task_id": "pipe_clamp_saddle",
3583
+ "code_chars": 4730,
3584
+ "cadforge_build": 0.0,
3585
+ "cadforge_total": -1.0,
3586
+ "has_fixture": true,
3587
+ "has_cadquery_import": true,
3588
+ "error_kind": "AttributeError"
3589
+ },
3590
+ {
3591
+ "reward_call": 75,
3592
+ "index": 3,
3593
+ "score": -0.615,
3594
+ "task_id": "pipe_clamp_saddle",
3595
+ "code_chars": 4192,
3596
+ "cadforge_build": 0.0,
3597
+ "cadforge_total": -1.0,
3598
+ "has_fixture": true,
3599
+ "has_cadquery_import": true,
3600
+ "error_kind": "NameError"
3601
+ },
3602
+ {
3603
+ "reward_call": 76,
3604
+ "index": 0,
3605
+ "score": 0.6222121869979141,
3606
+ "task_id": "wall_j_hook_120n",
3607
+ "code_chars": 2305,
3608
+ "cadforge_build": 1.0,
3609
+ "cadforge_total": 0.5555437494093108,
3610
+ "has_fixture": true,
3611
+ "has_cadquery_import": true,
3612
+ "error_kind": "build_ok"
3613
+ },
3614
+ {
3615
+ "reward_call": 76,
3616
+ "index": 1,
3617
+ "score": -0.565,
3618
+ "task_id": "wall_j_hook_120n",
3619
+ "code_chars": 3598,
3620
+ "cadforge_build": 0.0,
3621
+ "cadforge_total": -1.0,
3622
+ "has_fixture": true,
3623
+ "has_cadquery_import": true,
3624
+ "error_kind": "unknown"
3625
+ },
3626
+ {
3627
+ "reward_call": 76,
3628
+ "index": 2,
3629
+ "score": 0.7234843750000001,
3630
+ "task_id": "robot_gripper_two_finger",
3631
+ "code_chars": 3930,
3632
+ "cadforge_build": 1.0,
3633
+ "cadforge_total": 0.6746875000000001,
3634
+ "has_fixture": true,
3635
+ "has_cadquery_import": true,
3636
+ "error_kind": "build_ok"
3637
+ },
3638
+ {
3639
+ "reward_call": 76,
3640
+ "index": 3,
3641
+ "score": 0.7888723958333335,
3642
+ "task_id": "robot_gripper_two_finger",
3643
+ "code_chars": 3039,
3644
+ "cadforge_build": 1.0,
3645
+ "cadforge_total": 0.7516145833333334,
3646
+ "has_fixture": true,
3647
+ "has_cadquery_import": true,
3648
+ "error_kind": "build_ok"
3649
+ },
3650
+ {
3651
+ "reward_call": 77,
3652
+ "index": 0,
3653
+ "score": 0.8540833333333334,
3654
+ "task_id": "table_six_leg_500n",
3655
+ "code_chars": 5750,
3656
+ "cadforge_build": 1.0,
3657
+ "cadforge_total": 0.8283333333333334,
3658
+ "has_fixture": true,
3659
+ "has_cadquery_import": true,
3660
+ "error_kind": "build_ok"
3661
+ },
3662
+ {
3663
+ "reward_call": 77,
3664
+ "index": 1,
3665
+ "score": 0.8546711906033754,
3666
+ "task_id": "table_six_leg_500n",
3667
+ "code_chars": 3273,
3668
+ "cadforge_build": 1.0,
3669
+ "cadforge_total": 0.829024930121618,
3670
+ "has_fixture": true,
3671
+ "has_cadquery_import": true,
3672
+ "error_kind": "build_ok"
3673
+ },
3674
+ {
3675
+ "reward_call": 77,
3676
+ "index": 2,
3677
+ "score": -0.695,
3678
+ "task_id": "lightweight_truss_support_250n",
3679
+ "code_chars": 4610,
3680
+ "cadforge_build": 0.0,
3681
+ "cadforge_total": -1.0,
3682
+ "has_fixture": true,
3683
+ "has_cadquery_import": true,
3684
+ "error_kind": "SyntaxError"
3685
+ },
3686
+ {
3687
+ "reward_call": 77,
3688
+ "index": 3,
3689
+ "score": -0.565,
3690
+ "task_id": "lightweight_truss_support_250n",
3691
+ "code_chars": 4999,
3692
+ "cadforge_build": 0.0,
3693
+ "cadforge_total": -1.0,
3694
+ "has_fixture": true,
3695
+ "has_cadquery_import": true,
3696
+ "error_kind": "ValueError"
3697
+ },
3698
+ {
3699
+ "reward_call": 78,
3700
+ "index": 0,
3701
+ "score": 0.7214649946671322,
3702
+ "task_id": "shaft_torque_clamp_120nm",
3703
+ "code_chars": 4041,
3704
+ "cadforge_build": 1.0,
3705
+ "cadforge_total": 0.6723117584319203,
3706
+ "has_fixture": true,
3707
+ "has_cadquery_import": true,
3708
+ "error_kind": "build_ok"
3709
+ },
3710
+ {
3711
+ "reward_call": 78,
3712
+ "index": 1,
3713
+ "score": 0.6966119958133985,
3714
+ "task_id": "shaft_torque_clamp_120nm",
3715
+ "code_chars": 2833,
3716
+ "cadforge_build": 1.0,
3717
+ "cadforge_total": 0.6430729362510571,
3718
+ "has_fixture": true,
3719
+ "has_cadquery_import": true,
3720
+ "error_kind": "build_ok"
3721
+ },
3722
+ {
3723
+ "reward_call": 78,
3724
+ "index": 2,
3725
+ "score": -0.6649999999999999,
3726
+ "task_id": "bottle_cage_bike",
3727
+ "code_chars": 4702,
3728
+ "cadforge_build": 0.0,
3729
+ "cadforge_total": -1.0,
3730
+ "has_fixture": true,
3731
+ "has_cadquery_import": true,
3732
+ "error_kind": "SyntaxError"
3733
+ },
3734
+ {
3735
+ "reward_call": 78,
3736
+ "index": 3,
3737
+ "score": -0.615,
3738
+ "task_id": "bottle_cage_bike",
3739
+ "code_chars": 5633,
3740
+ "cadforge_build": 0.0,
3741
+ "cadforge_total": -1.0,
3742
+ "has_fixture": true,
3743
+ "has_cadquery_import": true,
3744
+ "error_kind": "NameError"
3745
+ },
3746
+ {
3747
+ "reward_call": 79,
3748
+ "index": 0,
3749
+ "score": 0.7213078184926907,
3750
+ "task_id": "cable_clip_snap",
3751
+ "code_chars": 3680,
3752
+ "cadforge_build": 1.0,
3753
+ "cadforge_total": 0.6721268452855184,
3754
+ "has_fixture": true,
3755
+ "has_cadquery_import": true,
3756
+ "error_kind": "build_ok"
3757
+ },
3758
+ {
3759
+ "reward_call": 79,
3760
+ "index": 1,
3761
+ "score": -0.615,
3762
+ "task_id": "cable_clip_snap",
3763
+ "code_chars": 5223,
3764
+ "cadforge_build": 0.0,
3765
+ "cadforge_total": -1.0,
3766
+ "has_fixture": true,
3767
+ "has_cadquery_import": true,
3768
+ "error_kind": "AttributeError"
3769
+ },
3770
+ {
3771
+ "reward_call": 79,
3772
+ "index": 2,
3773
+ "score": 0.9192936698617971,
3774
+ "task_id": "phone_stand_adjustable_hinge",
3775
+ "code_chars": 4126,
3776
+ "cadforge_build": 1.0,
3777
+ "cadforge_total": 0.9050513763079966,
3778
+ "has_fixture": true,
3779
+ "has_cadquery_import": true,
3780
+ "error_kind": "build_ok"
3781
+ },
3782
+ {
3783
+ "reward_call": 79,
3784
+ "index": 3,
3785
+ "score": 0.7875,
3786
+ "task_id": "phone_stand_adjustable_hinge",
3787
+ "code_chars": 3566,
3788
+ "cadforge_build": 1.0,
3789
+ "cadforge_total": 0.75,
3790
+ "has_fixture": true,
3791
+ "has_cadquery_import": true,
3792
+ "error_kind": "build_ok"
3793
+ },
3794
+ {
3795
+ "reward_call": 80,
3796
+ "index": 0,
3797
+ "score": 0.736593127816838,
3798
+ "task_id": "shelf_bracket_triangular_200n",
3799
+ "code_chars": 5101,
3800
+ "cadforge_build": 1.0,
3801
+ "cadforge_total": 0.6901095621374563,
3802
+ "has_fixture": true,
3803
+ "has_cadquery_import": true,
3804
+ "error_kind": "build_ok"
3805
+ },
3806
+ {
3807
+ "reward_call": 80,
3808
+ "index": 1,
3809
+ "score": -0.615,
3810
+ "task_id": "shelf_bracket_triangular_200n",
3811
+ "code_chars": 2796,
3812
+ "cadforge_build": 0.0,
3813
+ "cadforge_total": -1.0,
3814
+ "has_fixture": true,
3815
+ "has_cadquery_import": true,
3816
+ "error_kind": "AttributeError"
3817
+ },
3818
+ {
3819
+ "reward_call": 80,
3820
+ "index": 2,
3821
+ "score": -0.84,
3822
+ "task_id": "folding_step_stool",
3823
+ "code_chars": 5352,
3824
+ "cadforge_build": 0.0,
3825
+ "cadforge_total": -1.0,
3826
+ "has_fixture": false,
3827
+ "has_cadquery_import": true,
3828
+ "error_kind": "SyntaxError"
3829
+ },
3830
+ {
3831
+ "reward_call": 80,
3832
+ "index": 3,
3833
+ "score": -0.84,
3834
+ "task_id": "folding_step_stool",
3835
+ "code_chars": 5373,
3836
+ "cadforge_build": 0.0,
3837
+ "cadforge_total": -1.0,
3838
+ "has_fixture": false,
3839
+ "has_cadquery_import": true,
3840
+ "error_kind": "SyntaxError"
3841
+ }
3842
+ ]
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_code_health.png ADDED

Git LFS Details

  • SHA256: 4356320e9de40c11fffd88efd27439df3fa14f4c171917d4dfbf92a3eccbbe3a
  • Pointer size: 131 Bytes
  • Size of remote file: 284 kB
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_error_breakdown.png ADDED
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_reward_curve.png ADDED

Git LFS Details

  • SHA256: a6e820edbc58cb36512769fd40942dbcf10dc59f8a556ca5673aff5060f5983c
  • Pointer size: 131 Bytes
  • Size of remote file: 179 kB
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/metrics.json ADDED
@@ -0,0 +1,2009 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "loss": -0.0394,
4
+ "grad_norm": 0.5024,
5
+ "learning_rate": 3e-06,
6
+ "num_tokens": 8476.0,
7
+ "completions/mean_length": 1911.0,
8
+ "completions/min_length": 1561.0,
9
+ "completions/max_length": 2048.0,
10
+ "completions/clipped_ratio": 0.5,
11
+ "completions/mean_terminated_length": 1774.0,
12
+ "completions/min_terminated_length": 1561.0,
13
+ "completions/max_terminated_length": 1987.0,
14
+ "rewards/reward_func/mean": -0.3249,
15
+ "rewards/reward_func/std": 0.739,
16
+ "reward": -0.3249,
17
+ "reward_std": 0.534,
18
+ "frac_reward_zero_std": 0.0,
19
+ "entropy": 0.37,
20
+ "clip_ratio/low_mean": 0.0,
21
+ "clip_ratio/low_min": 0.0,
22
+ "clip_ratio/high_mean": 0.0,
23
+ "clip_ratio/high_max": 0.0,
24
+ "clip_ratio/region_mean": 0.0,
25
+ "epoch": 0.1
26
+ },
27
+ {
28
+ "loss": -0.01784,
29
+ "grad_norm": 0.5358,
30
+ "learning_rate": 2.963e-06,
31
+ "num_tokens": 16440.0,
32
+ "completions/mean_length": 1760.0,
33
+ "completions/min_length": 1535.0,
34
+ "completions/max_length": 2048.0,
35
+ "completions/clipped_ratio": 0.25,
36
+ "completions/mean_terminated_length": 1664.0,
37
+ "completions/min_terminated_length": 1535.0,
38
+ "completions/max_terminated_length": 1895.0,
39
+ "rewards/reward_func/mean": -0.2601,
40
+ "rewards/reward_func/std": 0.7651,
41
+ "reward": -0.2601,
42
+ "reward_std": 0.5231,
43
+ "frac_reward_zero_std": 0.0,
44
+ "entropy": 0.2786,
45
+ "clip_ratio/low_mean": 0.0,
46
+ "clip_ratio/low_min": 0.0,
47
+ "clip_ratio/high_mean": 0.0,
48
+ "clip_ratio/high_max": 0.0,
49
+ "clip_ratio/region_mean": 0.0,
50
+ "epoch": 0.2
51
+ },
52
+ {
53
+ "loss": 0.05911,
54
+ "grad_norm": 0.555,
55
+ "learning_rate": 2.925e-06,
56
+ "num_tokens": 24920.0,
57
+ "completions/mean_length": 1890.0,
58
+ "completions/min_length": 1416.0,
59
+ "completions/max_length": 2048.0,
60
+ "completions/clipped_ratio": 0.75,
61
+ "completions/mean_terminated_length": 1416.0,
62
+ "completions/min_terminated_length": 1416.0,
63
+ "completions/max_terminated_length": 1416.0,
64
+ "rewards/reward_func/mean": -0.3867,
65
+ "rewards/reward_func/std": 0.7942,
66
+ "reward": -0.3867,
67
+ "reward_std": 0.641,
68
+ "frac_reward_zero_std": 0.0,
69
+ "entropy": 0.264,
70
+ "clip_ratio/low_mean": 0.0,
71
+ "clip_ratio/low_min": 0.0,
72
+ "clip_ratio/high_mean": 0.0,
73
+ "clip_ratio/high_max": 0.0,
74
+ "clip_ratio/region_mean": 0.0,
75
+ "epoch": 0.3
76
+ },
77
+ {
78
+ "loss": 0.0,
79
+ "grad_norm": 0.3544,
80
+ "learning_rate": 2.888e-06,
81
+ "num_tokens": 33850.0,
82
+ "completions/mean_length": 2048.0,
83
+ "completions/min_length": 2048.0,
84
+ "completions/max_length": 2048.0,
85
+ "completions/clipped_ratio": 1.0,
86
+ "completions/mean_terminated_length": 0.0,
87
+ "completions/min_terminated_length": 0.0,
88
+ "completions/max_terminated_length": 0.0,
89
+ "rewards/reward_func/mean": -0.4657,
90
+ "rewards/reward_func/std": 0.7486,
91
+ "reward": -0.4657,
92
+ "reward_std": 0.5293,
93
+ "frac_reward_zero_std": 0.5,
94
+ "entropy": 0.2868,
95
+ "clip_ratio/low_mean": 0.0,
96
+ "clip_ratio/low_min": 0.0,
97
+ "clip_ratio/high_mean": 0.0,
98
+ "clip_ratio/high_max": 0.0,
99
+ "clip_ratio/region_mean": 0.0,
100
+ "epoch": 0.4
101
+ },
102
+ {
103
+ "loss": 0.03674,
104
+ "grad_norm": 0.5205,
105
+ "learning_rate": 2.85e-06,
106
+ "num_tokens": 42310.0,
107
+ "completions/mean_length": 1883.0,
108
+ "completions/min_length": 1522.0,
109
+ "completions/max_length": 2048.0,
110
+ "completions/clipped_ratio": 0.5,
111
+ "completions/mean_terminated_length": 1718.0,
112
+ "completions/min_terminated_length": 1522.0,
113
+ "completions/max_terminated_length": 1915.0,
114
+ "rewards/reward_func/mean": -0.7088,
115
+ "rewards/reward_func/std": 0.1033,
116
+ "reward": -0.7088,
117
+ "reward_std": 0.04773,
118
+ "frac_reward_zero_std": 0.0,
119
+ "entropy": 0.3428,
120
+ "clip_ratio/low_mean": 0.0,
121
+ "clip_ratio/low_min": 0.0,
122
+ "clip_ratio/high_mean": 0.0,
123
+ "clip_ratio/high_max": 0.0,
124
+ "clip_ratio/region_mean": 0.0,
125
+ "epoch": 0.5
126
+ },
127
+ {
128
+ "loss": 0.043,
129
+ "grad_norm": 0.5484,
130
+ "learning_rate": 2.813e-06,
131
+ "num_tokens": 50780.0,
132
+ "completions/mean_length": 1930.0,
133
+ "completions/min_length": 1578.0,
134
+ "completions/max_length": 2048.0,
135
+ "completions/clipped_ratio": 0.75,
136
+ "completions/mean_terminated_length": 1578.0,
137
+ "completions/min_terminated_length": 1578.0,
138
+ "completions/max_terminated_length": 1578.0,
139
+ "rewards/reward_func/mean": -0.8025,
140
+ "rewards/reward_func/std": 0.09242,
141
+ "reward": -0.8025,
142
+ "reward_std": 0.07071,
143
+ "frac_reward_zero_std": 0.0,
144
+ "entropy": 0.3512,
145
+ "clip_ratio/low_mean": 0.0,
146
+ "clip_ratio/low_min": 0.0,
147
+ "clip_ratio/high_mean": 0.0,
148
+ "clip_ratio/high_max": 0.0,
149
+ "clip_ratio/region_mean": 0.0,
150
+ "epoch": 0.6
151
+ },
152
+ {
153
+ "loss": 0.005404,
154
+ "grad_norm": 0.3593,
155
+ "learning_rate": 2.775e-06,
156
+ "num_tokens": 57960.0,
157
+ "completions/mean_length": 1570.0,
158
+ "completions/min_length": 1068.0,
159
+ "completions/max_length": 2048.0,
160
+ "completions/clipped_ratio": 0.5,
161
+ "completions/mean_terminated_length": 1092.0,
162
+ "completions/min_terminated_length": 1068.0,
163
+ "completions/max_terminated_length": 1116.0,
164
+ "rewards/reward_func/mean": -0.3876,
165
+ "rewards/reward_func/std": 0.733,
166
+ "reward": -0.3876,
167
+ "reward_std": 0.4453,
168
+ "frac_reward_zero_std": 0.5,
169
+ "entropy": 0.3426,
170
+ "clip_ratio/low_mean": 0.0,
171
+ "clip_ratio/low_min": 0.0,
172
+ "clip_ratio/high_mean": 0.0,
173
+ "clip_ratio/high_max": 0.0,
174
+ "clip_ratio/region_mean": 0.0,
175
+ "epoch": 0.7
176
+ },
177
+ {
178
+ "loss": 0.0427,
179
+ "grad_norm": 0.5516,
180
+ "learning_rate": 2.738e-06,
181
+ "num_tokens": 65940.0,
182
+ "completions/mean_length": 1776.0,
183
+ "completions/min_length": 1457.0,
184
+ "completions/max_length": 2048.0,
185
+ "completions/clipped_ratio": 0.25,
186
+ "completions/mean_terminated_length": 1685.0,
187
+ "completions/min_terminated_length": 1457.0,
188
+ "completions/max_terminated_length": 1880.0,
189
+ "rewards/reward_func/mean": -0.2678,
190
+ "rewards/reward_func/std": 0.7885,
191
+ "reward": -0.2678,
192
+ "reward_std": 0.6148,
193
+ "frac_reward_zero_std": 0.0,
194
+ "entropy": 0.2356,
195
+ "clip_ratio/low_mean": 0.0,
196
+ "clip_ratio/low_min": 0.0,
197
+ "clip_ratio/high_mean": 0.0,
198
+ "clip_ratio/high_max": 0.0,
199
+ "clip_ratio/region_mean": 0.0,
200
+ "epoch": 0.8
201
+ },
202
+ {
203
+ "loss": 0.07757,
204
+ "grad_norm": 0.651,
205
+ "learning_rate": 2.7e-06,
206
+ "num_tokens": 71860.0,
207
+ "completions/mean_length": 1242.0,
208
+ "completions/min_length": 689.0,
209
+ "completions/max_length": 1523.0,
210
+ "completions/clipped_ratio": 0.0,
211
+ "completions/mean_terminated_length": 1242.0,
212
+ "completions/min_terminated_length": 689.0,
213
+ "completions/max_terminated_length": 1523.0,
214
+ "rewards/reward_func/mean": -0.2699,
215
+ "rewards/reward_func/std": 0.6585,
216
+ "reward": -0.2699,
217
+ "reward_std": 0.488,
218
+ "frac_reward_zero_std": 0.0,
219
+ "entropy": 0.3257,
220
+ "clip_ratio/low_mean": 0.0,
221
+ "clip_ratio/low_min": 0.0,
222
+ "clip_ratio/high_mean": 0.0,
223
+ "clip_ratio/high_max": 0.0,
224
+ "clip_ratio/region_mean": 0.0,
225
+ "epoch": 0.9
226
+ },
227
+ {
228
+ "loss": 1.192e-07,
229
+ "grad_norm": 0.381,
230
+ "learning_rate": 2.662e-06,
231
+ "num_tokens": 80370.0,
232
+ "completions/mean_length": 1904.0,
233
+ "completions/min_length": 1526.0,
234
+ "completions/max_length": 2048.0,
235
+ "completions/clipped_ratio": 0.5,
236
+ "completions/mean_terminated_length": 1760.0,
237
+ "completions/min_terminated_length": 1526.0,
238
+ "completions/max_terminated_length": 1994.0,
239
+ "rewards/reward_func/mean": -0.6612,
240
+ "rewards/reward_func/std": 0.1195,
241
+ "reward": -0.6612,
242
+ "reward_std": 0.07955,
243
+ "frac_reward_zero_std": 0.5,
244
+ "entropy": 0.307,
245
+ "clip_ratio/low_mean": 0.0,
246
+ "clip_ratio/low_min": 0.0,
247
+ "clip_ratio/high_mean": 0.0,
248
+ "clip_ratio/high_max": 0.0,
249
+ "clip_ratio/region_mean": 0.0,
250
+ "epoch": 1.0
251
+ },
252
+ {
253
+ "loss": -0.003504,
254
+ "grad_norm": 0.5651,
255
+ "learning_rate": 2.625e-06,
256
+ "num_tokens": 88110.0,
257
+ "completions/mean_length": 1715.0,
258
+ "completions/min_length": 1365.0,
259
+ "completions/max_length": 2048.0,
260
+ "completions/clipped_ratio": 0.5,
261
+ "completions/mean_terminated_length": 1382.0,
262
+ "completions/min_terminated_length": 1365.0,
263
+ "completions/max_terminated_length": 1399.0,
264
+ "rewards/reward_func/mean": -0.3604,
265
+ "rewards/reward_func/std": 0.7789,
266
+ "reward": -0.3604,
267
+ "reward_std": 0.5157,
268
+ "frac_reward_zero_std": 0.0,
269
+ "entropy": 0.3672,
270
+ "clip_ratio/low_mean": 0.0,
271
+ "clip_ratio/low_min": 0.0,
272
+ "clip_ratio/high_mean": 0.0,
273
+ "clip_ratio/high_max": 0.0,
274
+ "clip_ratio/region_mean": 0.0,
275
+ "epoch": 1.1
276
+ },
277
+ {
278
+ "loss": 0.04245,
279
+ "grad_norm": 0.3754,
280
+ "learning_rate": 2.588e-06,
281
+ "num_tokens": 96650.0,
282
+ "completions/mean_length": 1932.0,
283
+ "completions/min_length": 1584.0,
284
+ "completions/max_length": 2048.0,
285
+ "completions/clipped_ratio": 0.75,
286
+ "completions/mean_terminated_length": 1584.0,
287
+ "completions/min_terminated_length": 1584.0,
288
+ "completions/max_terminated_length": 1584.0,
289
+ "rewards/reward_func/mean": -0.4243,
290
+ "rewards/reward_func/std": 0.8315,
291
+ "reward": -0.4243,
292
+ "reward_std": 0.5879,
293
+ "frac_reward_zero_std": 0.5,
294
+ "entropy": 0.298,
295
+ "clip_ratio/low_mean": 0.0,
296
+ "clip_ratio/low_min": 0.0,
297
+ "clip_ratio/high_mean": 0.0,
298
+ "clip_ratio/high_max": 0.0,
299
+ "clip_ratio/region_mean": 0.0,
300
+ "epoch": 1.2
301
+ },
302
+ {
303
+ "loss": 0.08523,
304
+ "grad_norm": 0.5713,
305
+ "learning_rate": 2.55e-06,
306
+ "num_tokens": 103700.0,
307
+ "completions/mean_length": 1541.0,
308
+ "completions/min_length": 946.0,
309
+ "completions/max_length": 2048.0,
310
+ "completions/clipped_ratio": 0.25,
311
+ "completions/mean_terminated_length": 1372.0,
312
+ "completions/min_terminated_length": 946.0,
313
+ "completions/max_terminated_length": 1764.0,
314
+ "rewards/reward_func/mean": -0.3645,
315
+ "rewards/reward_func/std": 0.6515,
316
+ "reward": -0.3645,
317
+ "reward_std": 0.5346,
318
+ "frac_reward_zero_std": 0.0,
319
+ "entropy": 0.3835,
320
+ "clip_ratio/low_mean": 0.0,
321
+ "clip_ratio/low_min": 0.0,
322
+ "clip_ratio/high_mean": 0.0,
323
+ "clip_ratio/high_max": 0.0,
324
+ "clip_ratio/region_mean": 0.0,
325
+ "epoch": 1.3
326
+ },
327
+ {
328
+ "loss": 0.04129,
329
+ "grad_norm": 0.3469,
330
+ "learning_rate": 2.513e-06,
331
+ "num_tokens": 112300.0,
332
+ "completions/mean_length": 1935.0,
333
+ "completions/min_length": 1596.0,
334
+ "completions/max_length": 2048.0,
335
+ "completions/clipped_ratio": 0.75,
336
+ "completions/mean_terminated_length": 1596.0,
337
+ "completions/min_terminated_length": 1596.0,
338
+ "completions/max_terminated_length": 1596.0,
339
+ "rewards/reward_func/mean": -0.4039,
340
+ "rewards/reward_func/std": 0.76,
341
+ "reward": -0.4039,
342
+ "reward_std": 0.493,
343
+ "frac_reward_zero_std": 0.5,
344
+ "entropy": 0.2683,
345
+ "clip_ratio/low_mean": 0.0,
346
+ "clip_ratio/low_min": 0.0,
347
+ "clip_ratio/high_mean": 0.0,
348
+ "clip_ratio/high_max": 0.0,
349
+ "clip_ratio/region_mean": 0.0,
350
+ "epoch": 1.4
351
+ },
352
+ {
353
+ "loss": 0.01495,
354
+ "grad_norm": 0.3619,
355
+ "learning_rate": 2.475e-06,
356
+ "num_tokens": 119500.0,
357
+ "completions/mean_length": 1588.0,
358
+ "completions/min_length": 1061.0,
359
+ "completions/max_length": 2048.0,
360
+ "completions/clipped_ratio": 0.5,
361
+ "completions/mean_terminated_length": 1128.0,
362
+ "completions/min_terminated_length": 1061.0,
363
+ "completions/max_terminated_length": 1196.0,
364
+ "rewards/reward_func/mean": -0.01555,
365
+ "rewards/reward_func/std": 0.9521,
366
+ "reward": -0.01555,
367
+ "reward_std": 0.01013,
368
+ "frac_reward_zero_std": 0.5,
369
+ "entropy": 0.3346,
370
+ "clip_ratio/low_mean": 0.0,
371
+ "clip_ratio/low_min": 0.0,
372
+ "clip_ratio/high_mean": 0.0,
373
+ "clip_ratio/high_max": 0.0,
374
+ "clip_ratio/region_mean": 0.0,
375
+ "epoch": 1.5
376
+ },
377
+ {
378
+ "loss": 0.05873,
379
+ "grad_norm": 0.6721,
380
+ "learning_rate": 2.437e-06,
381
+ "num_tokens": 125400.0,
382
+ "completions/mean_length": 1221.0,
383
+ "completions/min_length": 751.0,
384
+ "completions/max_length": 1490.0,
385
+ "completions/clipped_ratio": 0.0,
386
+ "completions/mean_terminated_length": 1221.0,
387
+ "completions/min_terminated_length": 751.0,
388
+ "completions/max_terminated_length": 1490.0,
389
+ "rewards/reward_func/mean": 0.3965,
390
+ "rewards/reward_func/std": 0.6622,
391
+ "reward": 0.3965,
392
+ "reward_std": 0.5071,
393
+ "frac_reward_zero_std": 0.0,
394
+ "entropy": 0.3414,
395
+ "clip_ratio/low_mean": 0.0,
396
+ "clip_ratio/low_min": 0.0,
397
+ "clip_ratio/high_mean": 0.0,
398
+ "clip_ratio/high_max": 0.0,
399
+ "clip_ratio/region_mean": 0.0,
400
+ "epoch": 1.6
401
+ },
402
+ {
403
+ "loss": 0.04682,
404
+ "grad_norm": 0.5606,
405
+ "learning_rate": 2.4e-06,
406
+ "num_tokens": 133000.0,
407
+ "completions/mean_length": 1706.0,
408
+ "completions/min_length": 1350.0,
409
+ "completions/max_length": 2048.0,
410
+ "completions/clipped_ratio": 0.25,
411
+ "completions/mean_terminated_length": 1593.0,
412
+ "completions/min_terminated_length": 1350.0,
413
+ "completions/max_terminated_length": 1837.0,
414
+ "rewards/reward_func/mean": 0.03911,
415
+ "rewards/reward_func/std": 0.8833,
416
+ "reward": 0.03911,
417
+ "reward_std": 1.07,
418
+ "frac_reward_zero_std": 0.0,
419
+ "entropy": 0.2752,
420
+ "clip_ratio/low_mean": 0.0,
421
+ "clip_ratio/low_min": 0.0,
422
+ "clip_ratio/high_mean": 0.0,
423
+ "clip_ratio/high_max": 0.0,
424
+ "clip_ratio/region_mean": 0.0,
425
+ "epoch": 1.7
426
+ },
427
+ {
428
+ "loss": -0.04875,
429
+ "grad_norm": 0.3378,
430
+ "learning_rate": 2.362e-06,
431
+ "num_tokens": 140800.0,
432
+ "completions/mean_length": 1718.0,
433
+ "completions/min_length": 1152.0,
434
+ "completions/max_length": 2048.0,
435
+ "completions/clipped_ratio": 0.5,
436
+ "completions/mean_terminated_length": 1389.0,
437
+ "completions/min_terminated_length": 1152.0,
438
+ "completions/max_terminated_length": 1626.0,
439
+ "rewards/reward_func/mean": -0.3733,
440
+ "rewards/reward_func/std": 0.7905,
441
+ "reward": -0.3733,
442
+ "reward_std": 0.5009,
443
+ "frac_reward_zero_std": 0.5,
444
+ "entropy": 0.3487,
445
+ "clip_ratio/low_mean": 0.0,
446
+ "clip_ratio/low_min": 0.0,
447
+ "clip_ratio/high_mean": 0.0,
448
+ "clip_ratio/high_max": 0.0,
449
+ "clip_ratio/region_mean": 0.0,
450
+ "epoch": 1.8
451
+ },
452
+ {
453
+ "loss": 0.07237,
454
+ "grad_norm": 0.3886,
455
+ "learning_rate": 2.325e-06,
456
+ "num_tokens": 148900.0,
457
+ "completions/mean_length": 1778.0,
458
+ "completions/min_length": 1228.0,
459
+ "completions/max_length": 1994.0,
460
+ "completions/clipped_ratio": 0.0,
461
+ "completions/mean_terminated_length": 1778.0,
462
+ "completions/min_terminated_length": 1228.0,
463
+ "completions/max_terminated_length": 1994.0,
464
+ "rewards/reward_func/mean": -0.3053,
465
+ "rewards/reward_func/std": 0.588,
466
+ "reward": -0.3053,
467
+ "reward_std": 0.438,
468
+ "frac_reward_zero_std": 0.5,
469
+ "entropy": 0.3341,
470
+ "clip_ratio/low_mean": 0.0,
471
+ "clip_ratio/low_min": 0.0,
472
+ "clip_ratio/high_mean": 0.0,
473
+ "clip_ratio/high_max": 0.0,
474
+ "clip_ratio/region_mean": 0.0,
475
+ "epoch": 1.9
476
+ },
477
+ {
478
+ "loss": 0.04621,
479
+ "grad_norm": 0.5304,
480
+ "learning_rate": 2.288e-06,
481
+ "num_tokens": 157200.0,
482
+ "completions/mean_length": 1890.0,
483
+ "completions/min_length": 1484.0,
484
+ "completions/max_length": 2048.0,
485
+ "completions/clipped_ratio": 0.5,
486
+ "completions/mean_terminated_length": 1731.0,
487
+ "completions/min_terminated_length": 1484.0,
488
+ "completions/max_terminated_length": 1978.0,
489
+ "rewards/reward_func/mean": -0.3186,
490
+ "rewards/reward_func/std": 0.7809,
491
+ "reward": -0.3186,
492
+ "reward_std": 0.5429,
493
+ "frac_reward_zero_std": 0.0,
494
+ "entropy": 0.3116,
495
+ "clip_ratio/low_mean": 0.0,
496
+ "clip_ratio/low_min": 0.0,
497
+ "clip_ratio/high_mean": 0.0,
498
+ "clip_ratio/high_max": 0.0,
499
+ "clip_ratio/region_mean": 0.0,
500
+ "epoch": 2.0
501
+ },
502
+ {
503
+ "loss": 0.02629,
504
+ "grad_norm": 0.5541,
505
+ "learning_rate": 2.25e-06,
506
+ "num_tokens": 164700.0,
507
+ "completions/mean_length": 1661.0,
508
+ "completions/min_length": 1241.0,
509
+ "completions/max_length": 1958.0,
510
+ "completions/clipped_ratio": 0.0,
511
+ "completions/mean_terminated_length": 1661.0,
512
+ "completions/min_terminated_length": 1241.0,
513
+ "completions/max_terminated_length": 1958.0,
514
+ "rewards/reward_func/mean": -0.276,
515
+ "rewards/reward_func/std": 0.6451,
516
+ "reward": -0.276,
517
+ "reward_std": 0.4794,
518
+ "frac_reward_zero_std": 0.0,
519
+ "entropy": 0.312,
520
+ "clip_ratio/low_mean": 0.0,
521
+ "clip_ratio/low_min": 0.0,
522
+ "clip_ratio/high_mean": 0.0,
523
+ "clip_ratio/high_max": 0.0,
524
+ "clip_ratio/region_mean": 0.0,
525
+ "epoch": 2.1
526
+ },
527
+ {
528
+ "loss": 0.07377,
529
+ "grad_norm": 0.356,
530
+ "learning_rate": 2.212e-06,
531
+ "num_tokens": 173000.0,
532
+ "completions/mean_length": 1854.0,
533
+ "completions/min_length": 1274.0,
534
+ "completions/max_length": 2048.0,
535
+ "completions/clipped_ratio": 0.75,
536
+ "completions/mean_terminated_length": 1274.0,
537
+ "completions/min_terminated_length": 1274.0,
538
+ "completions/max_terminated_length": 1274.0,
539
+ "rewards/reward_func/mean": -0.4335,
540
+ "rewards/reward_func/std": 0.8129,
541
+ "reward": -0.4335,
542
+ "reward_std": 0.5748,
543
+ "frac_reward_zero_std": 0.5,
544
+ "entropy": 0.2585,
545
+ "clip_ratio/low_mean": 0.0,
546
+ "clip_ratio/low_min": 0.0,
547
+ "clip_ratio/high_mean": 0.0,
548
+ "clip_ratio/high_max": 0.0,
549
+ "clip_ratio/region_mean": 0.0,
550
+ "epoch": 2.2
551
+ },
552
+ {
553
+ "loss": 0.1177,
554
+ "grad_norm": 0.5901,
555
+ "learning_rate": 2.175e-06,
556
+ "num_tokens": 180600.0,
557
+ "completions/mean_length": 1685.0,
558
+ "completions/min_length": 1039.0,
559
+ "completions/max_length": 2048.0,
560
+ "completions/clipped_ratio": 0.25,
561
+ "completions/mean_terminated_length": 1564.0,
562
+ "completions/min_terminated_length": 1039.0,
563
+ "completions/max_terminated_length": 1884.0,
564
+ "rewards/reward_func/mean": -0.3145,
565
+ "rewards/reward_func/std": 0.7596,
566
+ "reward": -0.3145,
567
+ "reward_std": 0.5487,
568
+ "frac_reward_zero_std": 0.0,
569
+ "entropy": 0.3193,
570
+ "clip_ratio/low_mean": 0.0,
571
+ "clip_ratio/low_min": 0.0,
572
+ "clip_ratio/high_mean": 0.0,
573
+ "clip_ratio/high_max": 0.0,
574
+ "clip_ratio/region_mean": 0.0,
575
+ "epoch": 2.3
576
+ },
577
+ {
578
+ "loss": 0.04592,
579
+ "grad_norm": 0.5826,
580
+ "learning_rate": 2.138e-06,
581
+ "num_tokens": 187900.0,
582
+ "completions/mean_length": 1624.0,
583
+ "completions/min_length": 992.0,
584
+ "completions/max_length": 2048.0,
585
+ "completions/clipped_ratio": 0.25,
586
+ "completions/mean_terminated_length": 1483.0,
587
+ "completions/min_terminated_length": 992.0,
588
+ "completions/max_terminated_length": 2046.0,
589
+ "rewards/reward_func/mean": -0.3496,
590
+ "rewards/reward_func/std": 0.7206,
591
+ "reward": -0.3496,
592
+ "reward_std": 0.5344,
593
+ "frac_reward_zero_std": 0.0,
594
+ "entropy": 0.3785,
595
+ "clip_ratio/low_mean": 0.0,
596
+ "clip_ratio/low_min": 0.0,
597
+ "clip_ratio/high_mean": 0.0,
598
+ "clip_ratio/high_max": 0.0,
599
+ "clip_ratio/region_mean": 0.0,
600
+ "epoch": 2.4
601
+ },
602
+ {
603
+ "loss": 0.1013,
604
+ "grad_norm": 0.3389,
605
+ "learning_rate": 2.1e-06,
606
+ "num_tokens": 195800.0,
607
+ "completions/mean_length": 1728.0,
608
+ "completions/min_length": 911.0,
609
+ "completions/max_length": 2048.0,
610
+ "completions/clipped_ratio": 0.5,
611
+ "completions/mean_terminated_length": 1408.0,
612
+ "completions/min_terminated_length": 911.0,
613
+ "completions/max_terminated_length": 1904.0,
614
+ "rewards/reward_func/mean": -0.01903,
615
+ "rewards/reward_func/std": 0.9482,
616
+ "reward": -0.01903,
617
+ "reward_std": 0.01753,
618
+ "frac_reward_zero_std": 0.5,
619
+ "entropy": 0.3162,
620
+ "clip_ratio/low_mean": 0.0,
621
+ "clip_ratio/low_min": 0.0,
622
+ "clip_ratio/high_mean": 0.0,
623
+ "clip_ratio/high_max": 0.0,
624
+ "clip_ratio/region_mean": 0.0,
625
+ "epoch": 2.5
626
+ },
627
+ {
628
+ "loss": 0.02564,
629
+ "grad_norm": 0.5532,
630
+ "learning_rate": 2.063e-06,
631
+ "num_tokens": 203500.0,
632
+ "completions/mean_length": 1704.0,
633
+ "completions/min_length": 1507.0,
634
+ "completions/max_length": 2025.0,
635
+ "completions/clipped_ratio": 0.0,
636
+ "completions/mean_terminated_length": 1704.0,
637
+ "completions/min_terminated_length": 1507.0,
638
+ "completions/max_terminated_length": 2025.0,
639
+ "rewards/reward_func/mean": -0.2326,
640
+ "rewards/reward_func/std": 0.6986,
641
+ "reward": -0.2326,
642
+ "reward_std": 0.5055,
643
+ "frac_reward_zero_std": 0.0,
644
+ "entropy": 0.2788,
645
+ "clip_ratio/low_mean": 0.0,
646
+ "clip_ratio/low_min": 0.0,
647
+ "clip_ratio/high_mean": 0.0,
648
+ "clip_ratio/high_max": 0.0,
649
+ "clip_ratio/region_mean": 0.0,
650
+ "epoch": 2.6
651
+ },
652
+ {
653
+ "loss": 0.02595,
654
+ "grad_norm": 0.5325,
655
+ "learning_rate": 2.025e-06,
656
+ "num_tokens": 212100.0,
657
+ "completions/mean_length": 1976.0,
658
+ "completions/min_length": 1758.0,
659
+ "completions/max_length": 2048.0,
660
+ "completions/clipped_ratio": 0.75,
661
+ "completions/mean_terminated_length": 1758.0,
662
+ "completions/min_terminated_length": 1758.0,
663
+ "completions/max_terminated_length": 1758.0,
664
+ "rewards/reward_func/mean": -0.4442,
665
+ "rewards/reward_func/std": 0.7469,
666
+ "reward": -0.4442,
667
+ "reward_std": 0.5809,
668
+ "frac_reward_zero_std": 0.0,
669
+ "entropy": 0.313,
670
+ "clip_ratio/low_mean": 0.0,
671
+ "clip_ratio/low_min": 0.0,
672
+ "clip_ratio/high_mean": 0.0,
673
+ "clip_ratio/high_max": 0.0,
674
+ "clip_ratio/region_mean": 0.0,
675
+ "epoch": 2.7
676
+ },
677
+ {
678
+ "loss": 0.0,
679
+ "grad_norm": 0.0,
680
+ "learning_rate": 1.988e-06,
681
+ "num_tokens": 221100.0,
682
+ "completions/mean_length": 2048.0,
683
+ "completions/min_length": 2048.0,
684
+ "completions/max_length": 2048.0,
685
+ "completions/clipped_ratio": 1.0,
686
+ "completions/mean_terminated_length": 0.0,
687
+ "completions/min_terminated_length": 0.0,
688
+ "completions/max_terminated_length": 0.0,
689
+ "rewards/reward_func/mean": -0.84,
690
+ "rewards/reward_func/std": 0.0,
691
+ "reward": -0.84,
692
+ "reward_std": 0.0,
693
+ "frac_reward_zero_std": 1.0,
694
+ "entropy": 0.3534,
695
+ "clip_ratio/low_mean": 0.0,
696
+ "clip_ratio/low_min": 0.0,
697
+ "clip_ratio/high_mean": 0.0,
698
+ "clip_ratio/high_max": 0.0,
699
+ "clip_ratio/region_mean": 0.0,
700
+ "epoch": 2.8
701
+ },
702
+ {
703
+ "loss": 0.1698,
704
+ "grad_norm": 0.7029,
705
+ "learning_rate": 1.95e-06,
706
+ "num_tokens": 227200.0,
707
+ "completions/mean_length": 1254.0,
708
+ "completions/min_length": 741.0,
709
+ "completions/max_length": 2048.0,
710
+ "completions/clipped_ratio": 0.25,
711
+ "completions/mean_terminated_length": 988.7,
712
+ "completions/min_terminated_length": 741.0,
713
+ "completions/max_terminated_length": 1161.0,
714
+ "rewards/reward_func/mean": -0.2931,
715
+ "rewards/reward_func/std": 0.6844,
716
+ "reward": -0.2931,
717
+ "reward_std": 0.4764,
718
+ "frac_reward_zero_std": 0.0,
719
+ "entropy": 0.3602,
720
+ "clip_ratio/low_mean": 0.0,
721
+ "clip_ratio/low_min": 0.0,
722
+ "clip_ratio/high_mean": 0.0,
723
+ "clip_ratio/high_max": 0.0,
724
+ "clip_ratio/region_mean": 0.0,
725
+ "epoch": 2.9
726
+ },
727
+ {
728
+ "loss": 0.02085,
729
+ "grad_norm": 0.5969,
730
+ "learning_rate": 1.912e-06,
731
+ "num_tokens": 233800.0,
732
+ "completions/mean_length": 1432.0,
733
+ "completions/min_length": 802.0,
734
+ "completions/max_length": 2048.0,
735
+ "completions/clipped_ratio": 0.25,
736
+ "completions/mean_terminated_length": 1227.0,
737
+ "completions/min_terminated_length": 802.0,
738
+ "completions/max_terminated_length": 1978.0,
739
+ "rewards/reward_func/mean": -0.3102,
740
+ "rewards/reward_func/std": 0.7988,
741
+ "reward": -0.3102,
742
+ "reward_std": 0.6255,
743
+ "frac_reward_zero_std": 0.0,
744
+ "entropy": 0.3085,
745
+ "clip_ratio/low_mean": 0.0,
746
+ "clip_ratio/low_min": 0.0,
747
+ "clip_ratio/high_mean": 0.0,
748
+ "clip_ratio/high_max": 0.0,
749
+ "clip_ratio/region_mean": 0.0,
750
+ "epoch": 3.0
751
+ },
752
+ {
753
+ "loss": -0.01515,
754
+ "grad_norm": 0.5645,
755
+ "learning_rate": 1.875e-06,
756
+ "num_tokens": 241700.0,
757
+ "completions/mean_length": 1786.0,
758
+ "completions/min_length": 1447.0,
759
+ "completions/max_length": 2048.0,
760
+ "completions/clipped_ratio": 0.25,
761
+ "completions/mean_terminated_length": 1698.0,
762
+ "completions/min_terminated_length": 1447.0,
763
+ "completions/max_terminated_length": 1986.0,
764
+ "rewards/reward_func/mean": -0.2485,
765
+ "rewards/reward_func/std": 0.7013,
766
+ "reward": -0.2485,
767
+ "reward_std": 0.5184,
768
+ "frac_reward_zero_std": 0.0,
769
+ "entropy": 0.2202,
770
+ "clip_ratio/low_mean": 0.0,
771
+ "clip_ratio/low_min": 0.0,
772
+ "clip_ratio/high_mean": 0.0,
773
+ "clip_ratio/high_max": 0.0,
774
+ "clip_ratio/region_mean": 0.0,
775
+ "epoch": 3.1
776
+ },
777
+ {
778
+ "loss": 0.1143,
779
+ "grad_norm": 0.5944,
780
+ "learning_rate": 1.838e-06,
781
+ "num_tokens": 248500.0,
782
+ "completions/mean_length": 1486.0,
783
+ "completions/min_length": 1021.0,
784
+ "completions/max_length": 2048.0,
785
+ "completions/clipped_ratio": 0.25,
786
+ "completions/mean_terminated_length": 1298.0,
787
+ "completions/min_terminated_length": 1021.0,
788
+ "completions/max_terminated_length": 1470.0,
789
+ "rewards/reward_func/mean": -0.6525,
790
+ "rewards/reward_func/std": 0.1264,
791
+ "reward": -0.6525,
792
+ "reward_std": 0.1025,
793
+ "frac_reward_zero_std": 0.0,
794
+ "entropy": 0.3867,
795
+ "clip_ratio/low_mean": 0.0,
796
+ "clip_ratio/low_min": 0.0,
797
+ "clip_ratio/high_mean": 0.0,
798
+ "clip_ratio/high_max": 0.0,
799
+ "clip_ratio/region_mean": 0.0,
800
+ "epoch": 3.2
801
+ },
802
+ {
803
+ "loss": 0.02104,
804
+ "grad_norm": 0.3771,
805
+ "learning_rate": 1.8e-06,
806
+ "num_tokens": 256300.0,
807
+ "completions/mean_length": 1730.0,
808
+ "completions/min_length": 1310.0,
809
+ "completions/max_length": 2048.0,
810
+ "completions/clipped_ratio": 0.5,
811
+ "completions/mean_terminated_length": 1413.0,
812
+ "completions/min_terminated_length": 1310.0,
813
+ "completions/max_terminated_length": 1516.0,
814
+ "rewards/reward_func/mean": -0.3425,
815
+ "rewards/reward_func/std": 0.822,
816
+ "reward": -0.3425,
817
+ "reward_std": 0.5092,
818
+ "frac_reward_zero_std": 0.5,
819
+ "entropy": 0.2264,
820
+ "clip_ratio/low_mean": 0.0,
821
+ "clip_ratio/low_min": 0.0,
822
+ "clip_ratio/high_mean": 0.0,
823
+ "clip_ratio/high_max": 0.0,
824
+ "clip_ratio/region_mean": 0.0,
825
+ "epoch": 3.3
826
+ },
827
+ {
828
+ "loss": 0.0613,
829
+ "grad_norm": 0.5263,
830
+ "learning_rate": 1.763e-06,
831
+ "num_tokens": 263000.0,
832
+ "completions/mean_length": 1460.0,
833
+ "completions/min_length": 947.0,
834
+ "completions/max_length": 2048.0,
835
+ "completions/clipped_ratio": 0.25,
836
+ "completions/mean_terminated_length": 1264.0,
837
+ "completions/min_terminated_length": 947.0,
838
+ "completions/max_terminated_length": 1720.0,
839
+ "rewards/reward_func/mean": -0.297,
840
+ "rewards/reward_func/std": 0.6566,
841
+ "reward": -0.297,
842
+ "reward_std": 0.485,
843
+ "frac_reward_zero_std": 0.0,
844
+ "entropy": 0.2879,
845
+ "clip_ratio/low_mean": 0.0,
846
+ "clip_ratio/low_min": 0.0,
847
+ "clip_ratio/high_mean": 0.0,
848
+ "clip_ratio/high_max": 0.0,
849
+ "clip_ratio/region_mean": 0.0,
850
+ "epoch": 3.4
851
+ },
852
+ {
853
+ "loss": -0.06794,
854
+ "grad_norm": 0.5798,
855
+ "learning_rate": 1.725e-06,
856
+ "num_tokens": 270600.0,
857
+ "completions/mean_length": 1677.0,
858
+ "completions/min_length": 1232.0,
859
+ "completions/max_length": 1860.0,
860
+ "completions/clipped_ratio": 0.0,
861
+ "completions/mean_terminated_length": 1677.0,
862
+ "completions/min_terminated_length": 1232.0,
863
+ "completions/max_terminated_length": 1860.0,
864
+ "rewards/reward_func/mean": 0.1559,
865
+ "rewards/reward_func/std": 0.8616,
866
+ "reward": 0.1559,
867
+ "reward_std": 0.03008,
868
+ "frac_reward_zero_std": 0.0,
869
+ "entropy": 0.2342,
870
+ "clip_ratio/low_mean": 0.0,
871
+ "clip_ratio/low_min": 0.0,
872
+ "clip_ratio/high_mean": 0.0,
873
+ "clip_ratio/high_max": 0.0,
874
+ "clip_ratio/region_mean": 0.0,
875
+ "epoch": 3.5
876
+ },
877
+ {
878
+ "loss": 0.03404,
879
+ "grad_norm": 0.3284,
880
+ "learning_rate": 1.687e-06,
881
+ "num_tokens": 278000.0,
882
+ "completions/mean_length": 1615.0,
883
+ "completions/min_length": 1026.0,
884
+ "completions/max_length": 2048.0,
885
+ "completions/clipped_ratio": 0.5,
886
+ "completions/mean_terminated_length": 1182.0,
887
+ "completions/min_terminated_length": 1026.0,
888
+ "completions/max_terminated_length": 1337.0,
889
+ "rewards/reward_func/mean": -0.3994,
890
+ "rewards/reward_func/std": 0.7388,
891
+ "reward": -0.3994,
892
+ "reward_std": 0.464,
893
+ "frac_reward_zero_std": 0.5,
894
+ "entropy": 0.3576,
895
+ "clip_ratio/low_mean": 0.0,
896
+ "clip_ratio/low_min": 0.0,
897
+ "clip_ratio/high_mean": 0.0,
898
+ "clip_ratio/high_max": 0.0,
899
+ "clip_ratio/region_mean": 0.0,
900
+ "epoch": 3.6
901
+ },
902
+ {
903
+ "loss": -0.07904,
904
+ "grad_norm": 0.6905,
905
+ "learning_rate": 1.65e-06,
906
+ "num_tokens": 284200.0,
907
+ "completions/mean_length": 1319.0,
908
+ "completions/min_length": 760.0,
909
+ "completions/max_length": 1994.0,
910
+ "completions/clipped_ratio": 0.0,
911
+ "completions/mean_terminated_length": 1319.0,
912
+ "completions/min_terminated_length": 760.0,
913
+ "completions/max_terminated_length": 1994.0,
914
+ "rewards/reward_func/mean": 0.05681,
915
+ "rewards/reward_func/std": 0.777,
916
+ "reward": 0.05681,
917
+ "reward_std": 0.05178,
918
+ "frac_reward_zero_std": 0.0,
919
+ "entropy": 0.3424,
920
+ "clip_ratio/low_mean": 0.0,
921
+ "clip_ratio/low_min": 0.0,
922
+ "clip_ratio/high_mean": 0.0,
923
+ "clip_ratio/high_max": 0.0,
924
+ "clip_ratio/region_mean": 0.0,
925
+ "epoch": 3.7
926
+ },
927
+ {
928
+ "loss": 0.0,
929
+ "grad_norm": 0.5112,
930
+ "learning_rate": 1.613e-06,
931
+ "num_tokens": 293100.0,
932
+ "completions/mean_length": 2048.0,
933
+ "completions/min_length": 2048.0,
934
+ "completions/max_length": 2048.0,
935
+ "completions/clipped_ratio": 1.0,
936
+ "completions/mean_terminated_length": 0.0,
937
+ "completions/min_terminated_length": 0.0,
938
+ "completions/max_terminated_length": 0.0,
939
+ "rewards/reward_func/mean": -0.7713,
940
+ "rewards/reward_func/std": 0.07465,
941
+ "reward": -0.7713,
942
+ "reward_std": 0.06187,
943
+ "frac_reward_zero_std": 0.0,
944
+ "entropy": 0.3519,
945
+ "clip_ratio/low_mean": 0.0,
946
+ "clip_ratio/low_min": 0.0,
947
+ "clip_ratio/high_mean": 0.0,
948
+ "clip_ratio/high_max": 0.0,
949
+ "clip_ratio/region_mean": 0.0,
950
+ "epoch": 3.8
951
+ },
952
+ {
953
+ "loss": 0.01405,
954
+ "grad_norm": 0.4741,
955
+ "learning_rate": 1.575e-06,
956
+ "num_tokens": 301300.0,
957
+ "completions/mean_length": 1823.0,
958
+ "completions/min_length": 1525.0,
959
+ "completions/max_length": 2048.0,
960
+ "completions/clipped_ratio": 0.5,
961
+ "completions/mean_terminated_length": 1598.0,
962
+ "completions/min_terminated_length": 1525.0,
963
+ "completions/max_terminated_length": 1670.0,
964
+ "rewards/reward_func/mean": -0.02206,
965
+ "rewards/reward_func/std": 0.8902,
966
+ "reward": -0.02206,
967
+ "reward_std": 0.09307,
968
+ "frac_reward_zero_std": 0.0,
969
+ "entropy": 0.3407,
970
+ "clip_ratio/low_mean": 0.0,
971
+ "clip_ratio/low_min": 0.0,
972
+ "clip_ratio/high_mean": 0.0,
973
+ "clip_ratio/high_max": 0.0,
974
+ "clip_ratio/region_mean": 0.0,
975
+ "epoch": 3.9
976
+ },
977
+ {
978
+ "loss": 0.04343,
979
+ "grad_norm": 0.5796,
980
+ "learning_rate": 1.537e-06,
981
+ "num_tokens": 308300.0,
982
+ "completions/mean_length": 1517.0,
983
+ "completions/min_length": 1190.0,
984
+ "completions/max_length": 2024.0,
985
+ "completions/clipped_ratio": 0.0,
986
+ "completions/mean_terminated_length": 1517.0,
987
+ "completions/min_terminated_length": 1190.0,
988
+ "completions/max_terminated_length": 2024.0,
989
+ "rewards/reward_func/mean": 0.03957,
990
+ "rewards/reward_func/std": 0.775,
991
+ "reward": 0.03957,
992
+ "reward_std": 0.04758,
993
+ "frac_reward_zero_std": 0.0,
994
+ "entropy": 0.2929,
995
+ "clip_ratio/low_mean": 0.0,
996
+ "clip_ratio/low_min": 0.0,
997
+ "clip_ratio/high_mean": 0.0,
998
+ "clip_ratio/high_max": 0.0,
999
+ "clip_ratio/region_mean": 0.0,
1000
+ "epoch": 4.0
1001
+ },
1002
+ {
1003
+ "loss": 0.003253,
1004
+ "grad_norm": 0.526,
1005
+ "learning_rate": 1.5e-06,
1006
+ "num_tokens": 313400.0,
1007
+ "completions/mean_length": 1087.0,
1008
+ "completions/min_length": 931.0,
1009
+ "completions/max_length": 1393.0,
1010
+ "completions/clipped_ratio": 0.0,
1011
+ "completions/mean_terminated_length": 1087.0,
1012
+ "completions/min_terminated_length": 931.0,
1013
+ "completions/max_terminated_length": 1393.0,
1014
+ "rewards/reward_func/mean": -0.2308,
1015
+ "rewards/reward_func/std": 0.7366,
1016
+ "reward": -0.2308,
1017
+ "reward_std": 0.5434,
1018
+ "frac_reward_zero_std": 0.5,
1019
+ "entropy": 0.3219,
1020
+ "clip_ratio/low_mean": 0.0,
1021
+ "clip_ratio/low_min": 0.0,
1022
+ "clip_ratio/high_mean": 0.0,
1023
+ "clip_ratio/high_max": 0.0,
1024
+ "clip_ratio/region_mean": 0.0,
1025
+ "epoch": 4.1
1026
+ },
1027
+ {
1028
+ "loss": -0.03596,
1029
+ "grad_norm": 0.637,
1030
+ "learning_rate": 1.462e-06,
1031
+ "num_tokens": 320500.0,
1032
+ "completions/mean_length": 1568.0,
1033
+ "completions/min_length": 1324.0,
1034
+ "completions/max_length": 1717.0,
1035
+ "completions/clipped_ratio": 0.0,
1036
+ "completions/mean_terminated_length": 1568.0,
1037
+ "completions/min_terminated_length": 1324.0,
1038
+ "completions/max_terminated_length": 1717.0,
1039
+ "rewards/reward_func/mean": -0.615,
1040
+ "rewards/reward_func/std": 0.04082,
1041
+ "reward": -0.615,
1042
+ "reward_std": 0.03536,
1043
+ "frac_reward_zero_std": 0.0,
1044
+ "entropy": 0.2742,
1045
+ "clip_ratio/low_mean": 0.0,
1046
+ "clip_ratio/low_min": 0.0,
1047
+ "clip_ratio/high_mean": 0.0,
1048
+ "clip_ratio/high_max": 0.0,
1049
+ "clip_ratio/region_mean": 0.0,
1050
+ "epoch": 4.2
1051
+ },
1052
+ {
1053
+ "loss": 0.04259,
1054
+ "grad_norm": 0.333,
1055
+ "learning_rate": 1.425e-06,
1056
+ "num_tokens": 329200.0,
1057
+ "completions/mean_length": 1932.0,
1058
+ "completions/min_length": 1582.0,
1059
+ "completions/max_length": 2048.0,
1060
+ "completions/clipped_ratio": 0.75,
1061
+ "completions/mean_terminated_length": 1582.0,
1062
+ "completions/min_terminated_length": 1582.0,
1063
+ "completions/max_terminated_length": 1582.0,
1064
+ "rewards/reward_func/mean": -0.7275,
1065
+ "rewards/reward_func/std": 0.1362,
1066
+ "reward": -0.7275,
1067
+ "reward_std": 0.03536,
1068
+ "frac_reward_zero_std": 0.5,
1069
+ "entropy": 0.3253,
1070
+ "clip_ratio/low_mean": 0.0,
1071
+ "clip_ratio/low_min": 0.0,
1072
+ "clip_ratio/high_mean": 0.0,
1073
+ "clip_ratio/high_max": 0.0,
1074
+ "clip_ratio/region_mean": 0.0,
1075
+ "epoch": 4.3
1076
+ },
1077
+ {
1078
+ "loss": 0.09913,
1079
+ "grad_norm": 0.5848,
1080
+ "learning_rate": 1.388e-06,
1081
+ "num_tokens": 335900.0,
1082
+ "completions/mean_length": 1467.0,
1083
+ "completions/min_length": 1021.0,
1084
+ "completions/max_length": 2048.0,
1085
+ "completions/clipped_ratio": 0.25,
1086
+ "completions/mean_terminated_length": 1273.0,
1087
+ "completions/min_terminated_length": 1021.0,
1088
+ "completions/max_terminated_length": 1501.0,
1089
+ "rewards/reward_func/mean": -0.3088,
1090
+ "rewards/reward_func/std": 0.7659,
1091
+ "reward": -0.3088,
1092
+ "reward_std": 0.5921,
1093
+ "frac_reward_zero_std": 0.0,
1094
+ "entropy": 0.3179,
1095
+ "clip_ratio/low_mean": 0.0,
1096
+ "clip_ratio/low_min": 0.0,
1097
+ "clip_ratio/high_mean": 0.0,
1098
+ "clip_ratio/high_max": 0.0,
1099
+ "clip_ratio/region_mean": 0.0,
1100
+ "epoch": 4.4
1101
+ },
1102
+ {
1103
+ "loss": -0.03226,
1104
+ "grad_norm": 0.4715,
1105
+ "learning_rate": 1.35e-06,
1106
+ "num_tokens": 344500.0,
1107
+ "completions/mean_length": 1932.0,
1108
+ "completions/min_length": 1639.0,
1109
+ "completions/max_length": 2048.0,
1110
+ "completions/clipped_ratio": 0.5,
1111
+ "completions/mean_terminated_length": 1816.0,
1112
+ "completions/min_terminated_length": 1639.0,
1113
+ "completions/max_terminated_length": 1992.0,
1114
+ "rewards/reward_func/mean": -0.06068,
1115
+ "rewards/reward_func/std": 0.8723,
1116
+ "reward": -0.06068,
1117
+ "reward_std": 0.05548,
1118
+ "frac_reward_zero_std": 0.0,
1119
+ "entropy": 0.3674,
1120
+ "clip_ratio/low_mean": 0.0,
1121
+ "clip_ratio/low_min": 0.0,
1122
+ "clip_ratio/high_mean": 0.0,
1123
+ "clip_ratio/high_max": 0.0,
1124
+ "clip_ratio/region_mean": 0.0,
1125
+ "epoch": 4.5
1126
+ },
1127
+ {
1128
+ "loss": 0.04084,
1129
+ "grad_norm": 0.566,
1130
+ "learning_rate": 1.313e-06,
1131
+ "num_tokens": 351600.0,
1132
+ "completions/mean_length": 1551.0,
1133
+ "completions/min_length": 989.0,
1134
+ "completions/max_length": 2048.0,
1135
+ "completions/clipped_ratio": 0.25,
1136
+ "completions/mean_terminated_length": 1385.0,
1137
+ "completions/min_terminated_length": 989.0,
1138
+ "completions/max_terminated_length": 1933.0,
1139
+ "rewards/reward_func/mean": -0.2945,
1140
+ "rewards/reward_func/std": 0.7639,
1141
+ "reward": -0.2945,
1142
+ "reward_std": 0.6088,
1143
+ "frac_reward_zero_std": 0.0,
1144
+ "entropy": 0.3682,
1145
+ "clip_ratio/low_mean": 0.0,
1146
+ "clip_ratio/low_min": 0.0,
1147
+ "clip_ratio/high_mean": 0.0,
1148
+ "clip_ratio/high_max": 0.0,
1149
+ "clip_ratio/region_mean": 0.0,
1150
+ "epoch": 4.6
1151
+ },
1152
+ {
1153
+ "loss": 0.06336,
1154
+ "grad_norm": 0.3676,
1155
+ "learning_rate": 1.275e-06,
1156
+ "num_tokens": 359700.0,
1157
+ "completions/mean_length": 1830.0,
1158
+ "completions/min_length": 1284.0,
1159
+ "completions/max_length": 2048.0,
1160
+ "completions/clipped_ratio": 0.5,
1161
+ "completions/mean_terminated_length": 1612.0,
1162
+ "completions/min_terminated_length": 1284.0,
1163
+ "completions/max_terminated_length": 1940.0,
1164
+ "rewards/reward_func/mean": -0.3988,
1165
+ "rewards/reward_func/std": 0.74,
1166
+ "reward": -0.3988,
1167
+ "reward_std": 0.4648,
1168
+ "frac_reward_zero_std": 0.5,
1169
+ "entropy": 0.2581,
1170
+ "clip_ratio/low_mean": 0.0,
1171
+ "clip_ratio/low_min": 0.0,
1172
+ "clip_ratio/high_mean": 0.0,
1173
+ "clip_ratio/high_max": 0.0,
1174
+ "clip_ratio/region_mean": 0.0,
1175
+ "epoch": 4.7
1176
+ },
1177
+ {
1178
+ "loss": -0.02457,
1179
+ "grad_norm": 0.65,
1180
+ "learning_rate": 1.238e-06,
1181
+ "num_tokens": 364100.0,
1182
+ "completions/mean_length": 856.2,
1183
+ "completions/min_length": 653.0,
1184
+ "completions/max_length": 1252.0,
1185
+ "completions/clipped_ratio": 0.0,
1186
+ "completions/mean_terminated_length": 856.2,
1187
+ "completions/min_terminated_length": 653.0,
1188
+ "completions/max_terminated_length": 1252.0,
1189
+ "rewards/reward_func/mean": 0.375,
1190
+ "rewards/reward_func/std": 0.6622,
1191
+ "reward": 0.375,
1192
+ "reward_std": 0.4935,
1193
+ "frac_reward_zero_std": 0.0,
1194
+ "entropy": 0.3359,
1195
+ "clip_ratio/low_mean": 0.0,
1196
+ "clip_ratio/low_min": 0.0,
1197
+ "clip_ratio/high_mean": 0.0,
1198
+ "clip_ratio/high_max": 0.0,
1199
+ "clip_ratio/region_mean": 0.0,
1200
+ "epoch": 4.8
1201
+ },
1202
+ {
1203
+ "loss": -0.05181,
1204
+ "grad_norm": 0.4089,
1205
+ "learning_rate": 1.2e-06,
1206
+ "num_tokens": 370600.0,
1207
+ "completions/mean_length": 1368.0,
1208
+ "completions/min_length": 835.0,
1209
+ "completions/max_length": 1923.0,
1210
+ "completions/clipped_ratio": 0.0,
1211
+ "completions/mean_terminated_length": 1368.0,
1212
+ "completions/min_terminated_length": 835.0,
1213
+ "completions/max_terminated_length": 1923.0,
1214
+ "rewards/reward_func/mean": -0.2594,
1215
+ "rewards/reward_func/std": 0.6796,
1216
+ "reward": -0.2594,
1217
+ "reward_std": 0.5029,
1218
+ "frac_reward_zero_std": 0.5,
1219
+ "entropy": 0.3041,
1220
+ "clip_ratio/low_mean": 0.0,
1221
+ "clip_ratio/low_min": 0.0,
1222
+ "clip_ratio/high_mean": 0.0,
1223
+ "clip_ratio/high_max": 0.0,
1224
+ "clip_ratio/region_mean": 0.0,
1225
+ "epoch": 4.9
1226
+ },
1227
+ {
1228
+ "loss": 0.02844,
1229
+ "grad_norm": 0.5206,
1230
+ "learning_rate": 1.163e-06,
1231
+ "num_tokens": 378200.0,
1232
+ "completions/mean_length": 1707.0,
1233
+ "completions/min_length": 1504.0,
1234
+ "completions/max_length": 2048.0,
1235
+ "completions/clipped_ratio": 0.25,
1236
+ "completions/mean_terminated_length": 1594.0,
1237
+ "completions/min_terminated_length": 1504.0,
1238
+ "completions/max_terminated_length": 1678.0,
1239
+ "rewards/reward_func/mean": -0.2996,
1240
+ "rewards/reward_func/std": 0.7891,
1241
+ "reward": -0.2996,
1242
+ "reward_std": 0.5698,
1243
+ "frac_reward_zero_std": 0.0,
1244
+ "entropy": 0.2356,
1245
+ "clip_ratio/low_mean": 0.0,
1246
+ "clip_ratio/low_min": 0.0,
1247
+ "clip_ratio/high_mean": 0.0,
1248
+ "clip_ratio/high_max": 0.0,
1249
+ "clip_ratio/region_mean": 0.0,
1250
+ "epoch": 5.0
1251
+ },
1252
+ {
1253
+ "loss": -0.02722,
1254
+ "grad_norm": 0.7234,
1255
+ "learning_rate": 1.125e-06,
1256
+ "num_tokens": 383600.0,
1257
+ "completions/mean_length": 1114.0,
1258
+ "completions/min_length": 980.0,
1259
+ "completions/max_length": 1334.0,
1260
+ "completions/clipped_ratio": 0.0,
1261
+ "completions/mean_terminated_length": 1114.0,
1262
+ "completions/min_terminated_length": 980.0,
1263
+ "completions/max_terminated_length": 1334.0,
1264
+ "rewards/reward_func/mean": -0.2694,
1265
+ "rewards/reward_func/std": 0.6787,
1266
+ "reward": -0.2694,
1267
+ "reward_std": 0.51,
1268
+ "frac_reward_zero_std": 0.0,
1269
+ "entropy": 0.3971,
1270
+ "clip_ratio/low_mean": 0.0,
1271
+ "clip_ratio/low_min": 0.0,
1272
+ "clip_ratio/high_mean": 0.0,
1273
+ "clip_ratio/high_max": 0.0,
1274
+ "clip_ratio/region_mean": 0.0,
1275
+ "epoch": 5.1
1276
+ },
1277
+ {
1278
+ "loss": 0.01329,
1279
+ "grad_norm": 0.5717,
1280
+ "learning_rate": 1.087e-06,
1281
+ "num_tokens": 390800.0,
1282
+ "completions/mean_length": 1622.0,
1283
+ "completions/min_length": 1257.0,
1284
+ "completions/max_length": 2048.0,
1285
+ "completions/clipped_ratio": 0.25,
1286
+ "completions/mean_terminated_length": 1480.0,
1287
+ "completions/min_terminated_length": 1257.0,
1288
+ "completions/max_terminated_length": 1924.0,
1289
+ "rewards/reward_func/mean": -0.6287,
1290
+ "rewards/reward_func/std": 0.1084,
1291
+ "reward": -0.6288,
1292
+ "reward_std": 0.09016,
1293
+ "frac_reward_zero_std": 0.0,
1294
+ "entropy": 0.3027,
1295
+ "clip_ratio/low_mean": 0.0,
1296
+ "clip_ratio/low_min": 0.0,
1297
+ "clip_ratio/high_mean": 0.0,
1298
+ "clip_ratio/high_max": 0.0,
1299
+ "clip_ratio/region_mean": 0.0,
1300
+ "epoch": 5.2
1301
+ },
1302
+ {
1303
+ "loss": 0.01083,
1304
+ "grad_norm": 0.5798,
1305
+ "learning_rate": 1.05e-06,
1306
+ "num_tokens": 398200.0,
1307
+ "completions/mean_length": 1592.0,
1308
+ "completions/min_length": 1321.0,
1309
+ "completions/max_length": 1814.0,
1310
+ "completions/clipped_ratio": 0.0,
1311
+ "completions/mean_terminated_length": 1592.0,
1312
+ "completions/min_terminated_length": 1321.0,
1313
+ "completions/max_terminated_length": 1814.0,
1314
+ "rewards/reward_func/mean": 0.3896,
1315
+ "rewards/reward_func/std": 0.692,
1316
+ "reward": 0.3896,
1317
+ "reward_std": 0.5046,
1318
+ "frac_reward_zero_std": 0.0,
1319
+ "entropy": 0.2924,
1320
+ "clip_ratio/low_mean": 0.0,
1321
+ "clip_ratio/low_min": 0.0,
1322
+ "clip_ratio/high_mean": 0.0,
1323
+ "clip_ratio/high_max": 0.0,
1324
+ "clip_ratio/region_mean": 0.0,
1325
+ "epoch": 5.3
1326
+ },
1327
+ {
1328
+ "loss": -0.03384,
1329
+ "grad_norm": 0.3456,
1330
+ "learning_rate": 1.013e-06,
1331
+ "num_tokens": 406100.0,
1332
+ "completions/mean_length": 1771.0,
1333
+ "completions/min_length": 1324.0,
1334
+ "completions/max_length": 2048.0,
1335
+ "completions/clipped_ratio": 0.5,
1336
+ "completions/mean_terminated_length": 1494.0,
1337
+ "completions/min_terminated_length": 1324.0,
1338
+ "completions/max_terminated_length": 1663.0,
1339
+ "rewards/reward_func/mean": -0.3746,
1340
+ "rewards/reward_func/std": 0.7585,
1341
+ "reward": -0.3746,
1342
+ "reward_std": 0.4637,
1343
+ "frac_reward_zero_std": 0.5,
1344
+ "entropy": 0.3841,
1345
+ "clip_ratio/low_mean": 0.0,
1346
+ "clip_ratio/low_min": 0.0,
1347
+ "clip_ratio/high_mean": 0.0,
1348
+ "clip_ratio/high_max": 0.0,
1349
+ "clip_ratio/region_mean": 0.0,
1350
+ "epoch": 5.4
1351
+ },
1352
+ {
1353
+ "loss": -0.009678,
1354
+ "grad_norm": 0.5577,
1355
+ "learning_rate": 9.75e-07,
1356
+ "num_tokens": 413700.0,
1357
+ "completions/mean_length": 1699.0,
1358
+ "completions/min_length": 1303.0,
1359
+ "completions/max_length": 2048.0,
1360
+ "completions/clipped_ratio": 0.5,
1361
+ "completions/mean_terminated_length": 1350.0,
1362
+ "completions/min_terminated_length": 1303.0,
1363
+ "completions/max_terminated_length": 1396.0,
1364
+ "rewards/reward_func/mean": -0.4064,
1365
+ "rewards/reward_func/std": 0.7171,
1366
+ "reward": -0.4064,
1367
+ "reward_std": 0.4399,
1368
+ "frac_reward_zero_std": 0.0,
1369
+ "entropy": 0.3817,
1370
+ "clip_ratio/low_mean": 0.0,
1371
+ "clip_ratio/low_min": 0.0,
1372
+ "clip_ratio/high_mean": 0.0,
1373
+ "clip_ratio/high_max": 0.0,
1374
+ "clip_ratio/region_mean": 0.0,
1375
+ "epoch": 5.5
1376
+ },
1377
+ {
1378
+ "loss": 0.02153,
1379
+ "grad_norm": 0.5283,
1380
+ "learning_rate": 9.375e-07,
1381
+ "num_tokens": 421100.0,
1382
+ "completions/mean_length": 1624.0,
1383
+ "completions/min_length": 1300.0,
1384
+ "completions/max_length": 2048.0,
1385
+ "completions/clipped_ratio": 0.25,
1386
+ "completions/mean_terminated_length": 1483.0,
1387
+ "completions/min_terminated_length": 1300.0,
1388
+ "completions/max_terminated_length": 1850.0,
1389
+ "rewards/reward_func/mean": 0.009537,
1390
+ "rewards/reward_func/std": 0.872,
1391
+ "reward": 0.009537,
1392
+ "reward_std": 0.06923,
1393
+ "frac_reward_zero_std": 0.0,
1394
+ "entropy": 0.3162,
1395
+ "clip_ratio/low_mean": 0.0,
1396
+ "clip_ratio/low_min": 0.0,
1397
+ "clip_ratio/high_mean": 0.0,
1398
+ "clip_ratio/high_max": 0.0,
1399
+ "clip_ratio/region_mean": 0.0,
1400
+ "epoch": 5.6
1401
+ },
1402
+ {
1403
+ "loss": 0.0186,
1404
+ "grad_norm": 0.5043,
1405
+ "learning_rate": 9e-07,
1406
+ "num_tokens": 429900.0,
1407
+ "completions/mean_length": 1996.0,
1408
+ "completions/min_length": 1838.0,
1409
+ "completions/max_length": 2048.0,
1410
+ "completions/clipped_ratio": 0.75,
1411
+ "completions/mean_terminated_length": 1838.0,
1412
+ "completions/min_terminated_length": 1838.0,
1413
+ "completions/max_terminated_length": 1838.0,
1414
+ "rewards/reward_func/mean": -0.4434,
1415
+ "rewards/reward_func/std": 0.7875,
1416
+ "reward": -0.4434,
1417
+ "reward_std": 0.5926,
1418
+ "frac_reward_zero_std": 0.0,
1419
+ "entropy": 0.3133,
1420
+ "clip_ratio/low_mean": 0.0,
1421
+ "clip_ratio/low_min": 0.0,
1422
+ "clip_ratio/high_mean": 0.0,
1423
+ "clip_ratio/high_max": 0.0,
1424
+ "clip_ratio/region_mean": 0.0,
1425
+ "epoch": 5.7
1426
+ },
1427
+ {
1428
+ "loss": -0.02743,
1429
+ "grad_norm": 0.493,
1430
+ "learning_rate": 8.625e-07,
1431
+ "num_tokens": 437900.0,
1432
+ "completions/mean_length": 1764.0,
1433
+ "completions/min_length": 1344.0,
1434
+ "completions/max_length": 2048.0,
1435
+ "completions/clipped_ratio": 0.5,
1436
+ "completions/mean_terminated_length": 1481.0,
1437
+ "completions/min_terminated_length": 1344.0,
1438
+ "completions/max_terminated_length": 1618.0,
1439
+ "rewards/reward_func/mean": 0.006844,
1440
+ "rewards/reward_func/std": 0.9237,
1441
+ "reward": 0.006844,
1442
+ "reward_std": 0.09589,
1443
+ "frac_reward_zero_std": 0.0,
1444
+ "entropy": 0.3011,
1445
+ "clip_ratio/low_mean": 0.0,
1446
+ "clip_ratio/low_min": 0.0,
1447
+ "clip_ratio/high_mean": 0.0,
1448
+ "clip_ratio/high_max": 0.0,
1449
+ "clip_ratio/region_mean": 0.0,
1450
+ "epoch": 5.8
1451
+ },
1452
+ {
1453
+ "loss": 0.09348,
1454
+ "grad_norm": 0.6156,
1455
+ "learning_rate": 8.25e-07,
1456
+ "num_tokens": 444700.0,
1457
+ "completions/mean_length": 1476.0,
1458
+ "completions/min_length": 1187.0,
1459
+ "completions/max_length": 2048.0,
1460
+ "completions/clipped_ratio": 0.25,
1461
+ "completions/mean_terminated_length": 1286.0,
1462
+ "completions/min_terminated_length": 1187.0,
1463
+ "completions/max_terminated_length": 1375.0,
1464
+ "rewards/reward_func/mean": -0.28,
1465
+ "rewards/reward_func/std": 0.705,
1466
+ "reward": -0.28,
1467
+ "reward_std": 0.5445,
1468
+ "frac_reward_zero_std": 0.0,
1469
+ "entropy": 0.3418,
1470
+ "clip_ratio/low_mean": 0.0,
1471
+ "clip_ratio/low_min": 0.0,
1472
+ "clip_ratio/high_mean": 0.0,
1473
+ "clip_ratio/high_max": 0.0,
1474
+ "clip_ratio/region_mean": 0.0,
1475
+ "epoch": 5.9
1476
+ },
1477
+ {
1478
+ "loss": 0.09161,
1479
+ "grad_norm": 0.6272,
1480
+ "learning_rate": 7.875e-07,
1481
+ "num_tokens": 451700.0,
1482
+ "completions/mean_length": 1548.0,
1483
+ "completions/min_length": 1088.0,
1484
+ "completions/max_length": 2048.0,
1485
+ "completions/clipped_ratio": 0.25,
1486
+ "completions/mean_terminated_length": 1381.0,
1487
+ "completions/min_terminated_length": 1088.0,
1488
+ "completions/max_terminated_length": 1606.0,
1489
+ "rewards/reward_func/mean": 0.04736,
1490
+ "rewards/reward_func/std": 0.7942,
1491
+ "reward": 0.04736,
1492
+ "reward_std": 0.9721,
1493
+ "frac_reward_zero_std": 0.0,
1494
+ "entropy": 0.334,
1495
+ "clip_ratio/low_mean": 0.0,
1496
+ "clip_ratio/low_min": 0.0,
1497
+ "clip_ratio/high_mean": 0.0,
1498
+ "clip_ratio/high_max": 0.0,
1499
+ "clip_ratio/region_mean": 0.0,
1500
+ "epoch": 6.0
1501
+ },
1502
+ {
1503
+ "loss": 0.008331,
1504
+ "grad_norm": 0.4663,
1505
+ "learning_rate": 7.5e-07,
1506
+ "num_tokens": 458300.0,
1507
+ "completions/mean_length": 1400.0,
1508
+ "completions/min_length": 723.0,
1509
+ "completions/max_length": 1730.0,
1510
+ "completions/clipped_ratio": 0.0,
1511
+ "completions/mean_terminated_length": 1400.0,
1512
+ "completions/min_terminated_length": 723.0,
1513
+ "completions/max_terminated_length": 1730.0,
1514
+ "rewards/reward_func/mean": -0.2456,
1515
+ "rewards/reward_func/std": 0.7354,
1516
+ "reward": -0.2456,
1517
+ "reward_std": 0.5188,
1518
+ "frac_reward_zero_std": 0.5,
1519
+ "entropy": 0.3412,
1520
+ "clip_ratio/low_mean": 0.0,
1521
+ "clip_ratio/low_min": 0.0,
1522
+ "clip_ratio/high_mean": 0.0,
1523
+ "clip_ratio/high_max": 0.0,
1524
+ "clip_ratio/region_mean": 0.0,
1525
+ "epoch": 6.1
1526
+ },
1527
+ {
1528
+ "loss": 0.03415,
1529
+ "grad_norm": 0.4937,
1530
+ "learning_rate": 7.125e-07,
1531
+ "num_tokens": 466800.0,
1532
+ "completions/mean_length": 1952.0,
1533
+ "completions/min_length": 1801.0,
1534
+ "completions/max_length": 2048.0,
1535
+ "completions/clipped_ratio": 0.5,
1536
+ "completions/mean_terminated_length": 1856.0,
1537
+ "completions/min_terminated_length": 1801.0,
1538
+ "completions/max_terminated_length": 1911.0,
1539
+ "rewards/reward_func/mean": -0.6825,
1540
+ "rewards/reward_func/std": 0.1146,
1541
+ "reward": -0.6825,
1542
+ "reward_std": 0.09899,
1543
+ "frac_reward_zero_std": 0.0,
1544
+ "entropy": 0.3288,
1545
+ "clip_ratio/low_mean": 0.0,
1546
+ "clip_ratio/low_min": 0.0,
1547
+ "clip_ratio/high_mean": 0.0,
1548
+ "clip_ratio/high_max": 0.0,
1549
+ "clip_ratio/region_mean": 0.0,
1550
+ "epoch": 6.2
1551
+ },
1552
+ {
1553
+ "loss": 0.07031,
1554
+ "grad_norm": 0.3227,
1555
+ "learning_rate": 6.75e-07,
1556
+ "num_tokens": 475000.0,
1557
+ "completions/mean_length": 1863.0,
1558
+ "completions/min_length": 1307.0,
1559
+ "completions/max_length": 2048.0,
1560
+ "completions/clipped_ratio": 0.75,
1561
+ "completions/mean_terminated_length": 1307.0,
1562
+ "completions/min_terminated_length": 1307.0,
1563
+ "completions/max_terminated_length": 1307.0,
1564
+ "rewards/reward_func/mean": -0.4081,
1565
+ "rewards/reward_func/std": 0.7517,
1566
+ "reward": -0.4081,
1567
+ "reward_std": 0.4871,
1568
+ "frac_reward_zero_std": 0.5,
1569
+ "entropy": 0.3444,
1570
+ "clip_ratio/low_mean": 0.0,
1571
+ "clip_ratio/low_min": 0.0,
1572
+ "clip_ratio/high_mean": 0.0,
1573
+ "clip_ratio/high_max": 0.0,
1574
+ "clip_ratio/region_mean": 0.0,
1575
+ "epoch": 6.3
1576
+ },
1577
+ {
1578
+ "loss": 0.0,
1579
+ "grad_norm": 0.3374,
1580
+ "learning_rate": 6.375e-07,
1581
+ "num_tokens": 483700.0,
1582
+ "completions/mean_length": 1936.0,
1583
+ "completions/min_length": 1686.0,
1584
+ "completions/max_length": 2048.0,
1585
+ "completions/clipped_ratio": 0.5,
1586
+ "completions/mean_terminated_length": 1824.0,
1587
+ "completions/min_terminated_length": 1686.0,
1588
+ "completions/max_terminated_length": 1961.0,
1589
+ "rewards/reward_func/mean": -0.6463,
1590
+ "rewards/reward_func/std": 0.0625,
1591
+ "reward": -0.6463,
1592
+ "reward_std": 0.04419,
1593
+ "frac_reward_zero_std": 0.5,
1594
+ "entropy": 0.334,
1595
+ "clip_ratio/low_mean": 0.0,
1596
+ "clip_ratio/low_min": 0.0,
1597
+ "clip_ratio/high_mean": 0.0,
1598
+ "clip_ratio/high_max": 0.0,
1599
+ "clip_ratio/region_mean": 0.0,
1600
+ "epoch": 6.4
1601
+ },
1602
+ {
1603
+ "loss": 0.1354,
1604
+ "grad_norm": 0.5798,
1605
+ "learning_rate": 6e-07,
1606
+ "num_tokens": 490900.0,
1607
+ "completions/mean_length": 1564.0,
1608
+ "completions/min_length": 1141.0,
1609
+ "completions/max_length": 2048.0,
1610
+ "completions/clipped_ratio": 0.25,
1611
+ "completions/mean_terminated_length": 1402.0,
1612
+ "completions/min_terminated_length": 1141.0,
1613
+ "completions/max_terminated_length": 1678.0,
1614
+ "rewards/reward_func/mean": 0.08063,
1615
+ "rewards/reward_func/std": 0.8786,
1616
+ "reward": 0.08063,
1617
+ "reward_std": 1.069,
1618
+ "frac_reward_zero_std": 0.0,
1619
+ "entropy": 0.2125,
1620
+ "clip_ratio/low_mean": 0.0,
1621
+ "clip_ratio/low_min": 0.0,
1622
+ "clip_ratio/high_mean": 0.0,
1623
+ "clip_ratio/high_max": 0.0,
1624
+ "clip_ratio/region_mean": 0.0,
1625
+ "epoch": 6.5
1626
+ },
1627
+ {
1628
+ "loss": -0.0281,
1629
+ "grad_norm": 0.523,
1630
+ "learning_rate": 5.625e-07,
1631
+ "num_tokens": 498400.0,
1632
+ "completions/mean_length": 1711.0,
1633
+ "completions/min_length": 1238.0,
1634
+ "completions/max_length": 2048.0,
1635
+ "completions/clipped_ratio": 0.25,
1636
+ "completions/mean_terminated_length": 1599.0,
1637
+ "completions/min_terminated_length": 1238.0,
1638
+ "completions/max_terminated_length": 2020.0,
1639
+ "rewards/reward_func/mean": -0.04604,
1640
+ "rewards/reward_func/std": 0.8337,
1641
+ "reward": -0.04604,
1642
+ "reward_std": 1.017,
1643
+ "frac_reward_zero_std": 0.0,
1644
+ "entropy": 0.2706,
1645
+ "clip_ratio/low_mean": 0.0,
1646
+ "clip_ratio/low_min": 0.0,
1647
+ "clip_ratio/high_mean": 0.0,
1648
+ "clip_ratio/high_max": 0.0,
1649
+ "clip_ratio/region_mean": 0.0,
1650
+ "epoch": 6.6
1651
+ },
1652
+ {
1653
+ "loss": 0.03199,
1654
+ "grad_norm": 0.7135,
1655
+ "learning_rate": 5.25e-07,
1656
+ "num_tokens": 504100.0,
1657
+ "completions/mean_length": 1204.0,
1658
+ "completions/min_length": 735.0,
1659
+ "completions/max_length": 1582.0,
1660
+ "completions/clipped_ratio": 0.0,
1661
+ "completions/mean_terminated_length": 1204.0,
1662
+ "completions/min_terminated_length": 735.0,
1663
+ "completions/max_terminated_length": 1582.0,
1664
+ "rewards/reward_func/mean": -0.2467,
1665
+ "rewards/reward_func/std": 0.6768,
1666
+ "reward": -0.2467,
1667
+ "reward_std": 0.4926,
1668
+ "frac_reward_zero_std": 0.0,
1669
+ "entropy": 0.3261,
1670
+ "clip_ratio/low_mean": 0.0,
1671
+ "clip_ratio/low_min": 0.0,
1672
+ "clip_ratio/high_mean": 0.0,
1673
+ "clip_ratio/high_max": 0.0,
1674
+ "clip_ratio/region_mean": 0.0,
1675
+ "epoch": 6.7
1676
+ },
1677
+ {
1678
+ "loss": -0.03352,
1679
+ "grad_norm": 0.6544,
1680
+ "learning_rate": 4.875e-07,
1681
+ "num_tokens": 510600.0,
1682
+ "completions/mean_length": 1385.0,
1683
+ "completions/min_length": 664.0,
1684
+ "completions/max_length": 1974.0,
1685
+ "completions/clipped_ratio": 0.0,
1686
+ "completions/mean_terminated_length": 1385.0,
1687
+ "completions/min_terminated_length": 664.0,
1688
+ "completions/max_terminated_length": 1974.0,
1689
+ "rewards/reward_func/mean": 0.07614,
1690
+ "rewards/reward_func/std": 0.7999,
1691
+ "reward": 0.07614,
1692
+ "reward_std": 0.06594,
1693
+ "frac_reward_zero_std": 0.0,
1694
+ "entropy": 0.3321,
1695
+ "clip_ratio/low_mean": 0.0,
1696
+ "clip_ratio/low_min": 0.0,
1697
+ "clip_ratio/high_mean": 0.0,
1698
+ "clip_ratio/high_max": 0.0,
1699
+ "clip_ratio/region_mean": 0.0,
1700
+ "epoch": 6.8
1701
+ },
1702
+ {
1703
+ "loss": 0.02528,
1704
+ "grad_norm": 0.5607,
1705
+ "learning_rate": 4.5e-07,
1706
+ "num_tokens": 518300.0,
1707
+ "completions/mean_length": 1692.0,
1708
+ "completions/min_length": 1215.0,
1709
+ "completions/max_length": 2048.0,
1710
+ "completions/clipped_ratio": 0.5,
1711
+ "completions/mean_terminated_length": 1336.0,
1712
+ "completions/min_terminated_length": 1215.0,
1713
+ "completions/max_terminated_length": 1457.0,
1714
+ "rewards/reward_func/mean": -0.3784,
1715
+ "rewards/reward_func/std": 0.7756,
1716
+ "reward": -0.3784,
1717
+ "reward_std": 0.483,
1718
+ "frac_reward_zero_std": 0.0,
1719
+ "entropy": 0.3514,
1720
+ "clip_ratio/low_mean": 0.0,
1721
+ "clip_ratio/low_min": 0.0,
1722
+ "clip_ratio/high_mean": 0.0,
1723
+ "clip_ratio/high_max": 0.0,
1724
+ "clip_ratio/region_mean": 0.0,
1725
+ "epoch": 6.9
1726
+ },
1727
+ {
1728
+ "loss": 0.06859,
1729
+ "grad_norm": 0.5341,
1730
+ "learning_rate": 4.125e-07,
1731
+ "num_tokens": 525900.0,
1732
+ "completions/mean_length": 1668.0,
1733
+ "completions/min_length": 964.0,
1734
+ "completions/max_length": 2048.0,
1735
+ "completions/clipped_ratio": 0.5,
1736
+ "completions/mean_terminated_length": 1288.0,
1737
+ "completions/min_terminated_length": 964.0,
1738
+ "completions/max_terminated_length": 1612.0,
1739
+ "rewards/reward_func/mean": -0.004183,
1740
+ "rewards/reward_func/std": 0.9377,
1741
+ "reward": -0.004183,
1742
+ "reward_std": 0.05967,
1743
+ "frac_reward_zero_std": 0.0,
1744
+ "entropy": 0.331,
1745
+ "clip_ratio/low_mean": 0.0,
1746
+ "clip_ratio/low_min": 0.0,
1747
+ "clip_ratio/high_mean": 0.0,
1748
+ "clip_ratio/high_max": 0.0,
1749
+ "clip_ratio/region_mean": 0.0,
1750
+ "epoch": 7.0
1751
+ },
1752
+ {
1753
+ "loss": 0.05499,
1754
+ "grad_norm": 0.6775,
1755
+ "learning_rate": 3.75e-07,
1756
+ "num_tokens": 531800.0,
1757
+ "completions/mean_length": 1261.0,
1758
+ "completions/min_length": 841.0,
1759
+ "completions/max_length": 1783.0,
1760
+ "completions/clipped_ratio": 0.0,
1761
+ "completions/mean_terminated_length": 1261.0,
1762
+ "completions/min_terminated_length": 841.0,
1763
+ "completions/max_terminated_length": 1783.0,
1764
+ "rewards/reward_func/mean": -0.2873,
1765
+ "rewards/reward_func/std": 0.6768,
1766
+ "reward": -0.2873,
1767
+ "reward_std": 0.4634,
1768
+ "frac_reward_zero_std": 0.0,
1769
+ "entropy": 0.3011,
1770
+ "clip_ratio/low_mean": 0.0,
1771
+ "clip_ratio/low_min": 0.0,
1772
+ "clip_ratio/high_mean": 0.0,
1773
+ "clip_ratio/high_max": 0.0,
1774
+ "clip_ratio/region_mean": 0.0,
1775
+ "epoch": 7.1
1776
+ },
1777
+ {
1778
+ "loss": 0.04002,
1779
+ "grad_norm": 0.4042,
1780
+ "learning_rate": 3.375e-07,
1781
+ "num_tokens": 539600.0,
1782
+ "completions/mean_length": 1696.0,
1783
+ "completions/min_length": 1152.0,
1784
+ "completions/max_length": 2048.0,
1785
+ "completions/clipped_ratio": 0.5,
1786
+ "completions/mean_terminated_length": 1344.0,
1787
+ "completions/min_terminated_length": 1152.0,
1788
+ "completions/max_terminated_length": 1536.0,
1789
+ "rewards/reward_func/mean": -0.358,
1790
+ "rewards/reward_func/std": 0.8209,
1791
+ "reward": -0.358,
1792
+ "reward_std": 0.5225,
1793
+ "frac_reward_zero_std": 0.5,
1794
+ "entropy": 0.3425,
1795
+ "clip_ratio/low_mean": 0.0,
1796
+ "clip_ratio/low_min": 0.0,
1797
+ "clip_ratio/high_mean": 0.0,
1798
+ "clip_ratio/high_max": 0.0,
1799
+ "clip_ratio/region_mean": 0.0,
1800
+ "epoch": 7.2
1801
+ },
1802
+ {
1803
+ "loss": -0.00619,
1804
+ "grad_norm": 0.5428,
1805
+ "learning_rate": 3e-07,
1806
+ "num_tokens": 548000.0,
1807
+ "completions/mean_length": 1886.0,
1808
+ "completions/min_length": 1690.0,
1809
+ "completions/max_length": 2048.0,
1810
+ "completions/clipped_ratio": 0.25,
1811
+ "completions/mean_terminated_length": 1831.0,
1812
+ "completions/min_terminated_length": 1690.0,
1813
+ "completions/max_terminated_length": 2002.0,
1814
+ "rewards/reward_func/mean": -0.3142,
1815
+ "rewards/reward_func/std": 0.7602,
1816
+ "reward": -0.3142,
1817
+ "reward_std": 0.5492,
1818
+ "frac_reward_zero_std": 0.0,
1819
+ "entropy": 0.3456,
1820
+ "clip_ratio/low_mean": 0.0,
1821
+ "clip_ratio/low_min": 0.0,
1822
+ "clip_ratio/high_mean": 0.0,
1823
+ "clip_ratio/high_max": 0.0,
1824
+ "clip_ratio/region_mean": 0.0,
1825
+ "epoch": 7.3
1826
+ },
1827
+ {
1828
+ "loss": 0.003039,
1829
+ "grad_norm": 0.4069,
1830
+ "learning_rate": 2.625e-07,
1831
+ "num_tokens": 555300.0,
1832
+ "completions/mean_length": 1628.0,
1833
+ "completions/min_length": 1217.0,
1834
+ "completions/max_length": 2048.0,
1835
+ "completions/clipped_ratio": 0.25,
1836
+ "completions/mean_terminated_length": 1489.0,
1837
+ "completions/min_terminated_length": 1217.0,
1838
+ "completions/max_terminated_length": 2020.0,
1839
+ "rewards/reward_func/mean": -0.3228,
1840
+ "rewards/reward_func/std": 0.7172,
1841
+ "reward": -0.3228,
1842
+ "reward_std": 0.5582,
1843
+ "frac_reward_zero_std": 0.5,
1844
+ "entropy": 0.342,
1845
+ "clip_ratio/low_mean": 0.0,
1846
+ "clip_ratio/low_min": 0.0,
1847
+ "clip_ratio/high_mean": 0.0,
1848
+ "clip_ratio/high_max": 0.0,
1849
+ "clip_ratio/region_mean": 0.0,
1850
+ "epoch": 7.4
1851
+ },
1852
+ {
1853
+ "loss": 0.0276,
1854
+ "grad_norm": 0.5313,
1855
+ "learning_rate": 2.25e-07,
1856
+ "num_tokens": 563200.0,
1857
+ "completions/mean_length": 1785.0,
1858
+ "completions/min_length": 1392.0,
1859
+ "completions/max_length": 2048.0,
1860
+ "completions/clipped_ratio": 0.25,
1861
+ "completions/mean_terminated_length": 1697.0,
1862
+ "completions/min_terminated_length": 1392.0,
1863
+ "completions/max_terminated_length": 2038.0,
1864
+ "rewards/reward_func/mean": -0.3184,
1865
+ "rewards/reward_func/std": 0.7697,
1866
+ "reward": -0.3184,
1867
+ "reward_std": 0.5998,
1868
+ "frac_reward_zero_std": 0.0,
1869
+ "entropy": 0.3686,
1870
+ "clip_ratio/low_mean": 0.0,
1871
+ "clip_ratio/low_min": 0.0,
1872
+ "clip_ratio/high_mean": 0.0,
1873
+ "clip_ratio/high_max": 0.0,
1874
+ "clip_ratio/region_mean": 0.0,
1875
+ "epoch": 7.5
1876
+ },
1877
+ {
1878
+ "loss": 0.1387,
1879
+ "grad_norm": 0.6703,
1880
+ "learning_rate": 1.875e-07,
1881
+ "num_tokens": 568600.0,
1882
+ "completions/mean_length": 1112.0,
1883
+ "completions/min_length": 775.0,
1884
+ "completions/max_length": 1488.0,
1885
+ "completions/clipped_ratio": 0.0,
1886
+ "completions/mean_terminated_length": 1112.0,
1887
+ "completions/min_terminated_length": 775.0,
1888
+ "completions/max_terminated_length": 1488.0,
1889
+ "rewards/reward_func/mean": 0.3924,
1890
+ "rewards/reward_func/std": 0.6419,
1891
+ "reward": 0.3924,
1892
+ "reward_std": 0.4429,
1893
+ "frac_reward_zero_std": 0.0,
1894
+ "entropy": 0.2503,
1895
+ "clip_ratio/low_mean": 0.0,
1896
+ "clip_ratio/low_min": 0.0,
1897
+ "clip_ratio/high_mean": 0.0,
1898
+ "clip_ratio/high_max": 0.0,
1899
+ "clip_ratio/region_mean": 0.0,
1900
+ "epoch": 7.6
1901
+ },
1902
+ {
1903
+ "loss": 0.09392,
1904
+ "grad_norm": 0.4649,
1905
+ "learning_rate": 1.5e-07,
1906
+ "num_tokens": 576100.0,
1907
+ "completions/mean_length": 1666.0,
1908
+ "completions/min_length": 995.0,
1909
+ "completions/max_length": 2048.0,
1910
+ "completions/clipped_ratio": 0.25,
1911
+ "completions/mean_terminated_length": 1538.0,
1912
+ "completions/min_terminated_length": 995.0,
1913
+ "completions/max_terminated_length": 1815.0,
1914
+ "rewards/reward_func/mean": 0.1122,
1915
+ "rewards/reward_func/std": 0.8586,
1916
+ "reward": 0.1122,
1917
+ "reward_std": 0.04617,
1918
+ "frac_reward_zero_std": 0.0,
1919
+ "entropy": 0.3404,
1920
+ "clip_ratio/low_mean": 0.0,
1921
+ "clip_ratio/low_min": 0.0,
1922
+ "clip_ratio/high_mean": 0.0,
1923
+ "clip_ratio/high_max": 0.0,
1924
+ "clip_ratio/region_mean": 0.0,
1925
+ "epoch": 7.7
1926
+ },
1927
+ {
1928
+ "loss": -0.06634,
1929
+ "grad_norm": 0.593,
1930
+ "learning_rate": 1.125e-07,
1931
+ "num_tokens": 583200.0,
1932
+ "completions/mean_length": 1538.0,
1933
+ "completions/min_length": 1009.0,
1934
+ "completions/max_length": 1885.0,
1935
+ "completions/clipped_ratio": 0.0,
1936
+ "completions/mean_terminated_length": 1538.0,
1937
+ "completions/min_terminated_length": 1009.0,
1938
+ "completions/max_terminated_length": 1885.0,
1939
+ "rewards/reward_func/mean": 0.03452,
1940
+ "rewards/reward_func/std": 0.7792,
1941
+ "reward": 0.03452,
1942
+ "reward_std": 0.02646,
1943
+ "frac_reward_zero_std": 0.0,
1944
+ "entropy": 0.4021,
1945
+ "clip_ratio/low_mean": 0.0,
1946
+ "clip_ratio/low_min": 0.0,
1947
+ "clip_ratio/high_mean": 0.0,
1948
+ "clip_ratio/high_max": 0.0,
1949
+ "clip_ratio/region_mean": 0.0,
1950
+ "epoch": 7.8
1951
+ },
1952
+ {
1953
+ "loss": 0.0535,
1954
+ "grad_norm": 0.6048,
1955
+ "learning_rate": 7.5e-08,
1956
+ "num_tokens": 589700.0,
1957
+ "completions/mean_length": 1421.0,
1958
+ "completions/min_length": 1258.0,
1959
+ "completions/max_length": 1799.0,
1960
+ "completions/clipped_ratio": 0.0,
1961
+ "completions/mean_terminated_length": 1421.0,
1962
+ "completions/min_terminated_length": 1258.0,
1963
+ "completions/max_terminated_length": 1799.0,
1964
+ "rewards/reward_func/mean": 0.4533,
1965
+ "rewards/reward_func/std": 0.7169,
1966
+ "reward": 0.4533,
1967
+ "reward_std": 0.5191,
1968
+ "frac_reward_zero_std": 0.0,
1969
+ "entropy": 0.2737,
1970
+ "clip_ratio/low_mean": 0.0,
1971
+ "clip_ratio/low_min": 0.0,
1972
+ "clip_ratio/high_mean": 0.0,
1973
+ "clip_ratio/high_max": 0.0,
1974
+ "clip_ratio/region_mean": 0.0,
1975
+ "epoch": 7.9
1976
+ },
1977
+ {
1978
+ "loss": -0.06493,
1979
+ "grad_norm": 0.3786,
1980
+ "learning_rate": 3.75e-08,
1981
+ "num_tokens": 597100.0,
1982
+ "completions/mean_length": 1609.0,
1983
+ "completions/min_length": 874.0,
1984
+ "completions/max_length": 2048.0,
1985
+ "completions/clipped_ratio": 0.5,
1986
+ "completions/mean_terminated_length": 1170.0,
1987
+ "completions/min_terminated_length": 874.0,
1988
+ "completions/max_terminated_length": 1465.0,
1989
+ "rewards/reward_func/mean": -0.3896,
1990
+ "rewards/reward_func/std": 0.7583,
1991
+ "reward": -0.3896,
1992
+ "reward_std": 0.4779,
1993
+ "frac_reward_zero_std": 0.5,
1994
+ "entropy": 0.4093,
1995
+ "clip_ratio/low_mean": 0.0,
1996
+ "clip_ratio/low_min": 0.0,
1997
+ "clip_ratio/high_mean": 0.0,
1998
+ "clip_ratio/high_max": 0.0,
1999
+ "clip_ratio/region_mean": 0.0,
2000
+ "epoch": 8.0
2001
+ },
2002
+ {
2003
+ "train_runtime": 9112.0,
2004
+ "train_samples_per_second": 0.035,
2005
+ "train_steps_per_second": 0.009,
2006
+ "train_loss": 0.02709,
2007
+ "epoch": 8.0
2008
+ }
2009
+ ]
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/optimizer_health.png ADDED

Git LFS Details

  • SHA256: 709a605e6670e623db5f1ee7780b7a7087d56aff9eb7cfd6c9985ae3bcae07f8
  • Pointer size: 131 Bytes
  • Size of remote file: 145 kB
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/sft_loss_curve.png ADDED

Git LFS Details

  • SHA256: ddab2ef7e72cbbc694b408a35900bb324ec449b0e2971f34c65e81734bdcf2a6
  • Pointer size: 131 Bytes
  • Size of remote file: 135 kB
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/tensorboard_scalars.json ADDED
The diff for this file is too large to render. See raw diff
 
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/training_curve_report.md ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # CADForge Training Curves
2
+
3
+ - Log: `training/logs/grpo-9b-strict-build-20260426-strict-build.log`
4
+ - Parsed log metric rows: `81`
5
+ - TensorBoard scalar tags: `28`
6
+ - Train steps logged: `80`
7
+ - Eval points logged: `0`
8
+ - Train loss: `-0.0394` -> `-0.0649`
9
+ - GRPO reward points: `80`
10
+ - GRPO mean/best reward: `-0.2616` / `0.4533`
11
+ - GRPO trend slope: `0.003549` per step
12
+ - Completion debug rows: `320`
13
+ - Build rate from debug rows: `30.0%`
14
+ - Final fixture/import rates: `72.5%` / `99.7%`
15
+
16
+ ## Charts
17
+
18
+ ![sft_loss_curve](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/sft_loss_curve.png)
19
+ ![training_timeline](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/training_timeline.png)
20
+ ![optimizer_health](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/optimizer_health.png)
21
+ ![grpo_reward_curve](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_reward_curve.png)
22
+ ![grpo_code_health](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_code_health.png)
23
+ ![grpo_error_breakdown](training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/grpo_error_breakdown.png)
24
+
25
+ ## Final Trainer Metrics
26
+
27
+ ```json
28
+ {
29
+ "train_runtime": 9112.0,
30
+ "train_samples_per_second": 0.035,
31
+ "train_steps_per_second": 0.009,
32
+ "train_loss": 0.02709,
33
+ "epoch": 8.0
34
+ }
35
+ ```
36
+
37
+ ## Recent Train Loss
38
+
39
+ | step | loss |
40
+ |---:|---:|
41
+ | 61 | 0.0083 |
42
+ | 62 | 0.0341 |
43
+ | 63 | 0.0703 |
44
+ | 64 | 0.0000 |
45
+ | 65 | 0.1354 |
46
+ | 66 | -0.0281 |
47
+ | 67 | 0.0320 |
48
+ | 68 | -0.0335 |
49
+ | 69 | 0.0253 |
50
+ | 70 | 0.0686 |
51
+ | 71 | 0.0550 |
52
+ | 72 | 0.0400 |
53
+ | 73 | -0.0062 |
54
+ | 74 | 0.0030 |
55
+ | 75 | 0.0276 |
56
+ | 76 | 0.1387 |
57
+ | 77 | 0.0939 |
58
+ | 78 | -0.0663 |
59
+ | 79 | 0.0535 |
60
+ | 80 | -0.0649 |
61
+
62
+ ## GRPO Error Breakdown
63
+
64
+ | outcome | count |
65
+ |---|---:|
66
+ | `SyntaxError` | 109 |
67
+ | `build_ok` | 96 |
68
+ | `TypeError` | 25 |
69
+ | `ValueError` | 24 |
70
+ | `NameError` | 24 |
71
+ | `AttributeError` | 21 |
72
+ | `unknown` | 15 |
73
+ | `BRep_API` | 5 |
74
+ | `StdFail` | 1 |
training/reports/qwen35-9b-grpo-strict-build-20260426-strict-build/training_timeline.png ADDED