training: 200 steps final run, lr=1e-5

train/job_train.py

@@ -247,15 +247,13 @@ PROFILES: dict[str, dict] = {
         #                           stays flat for 50 consecutive steps.
         "grpo_lr":         "1e-5",
         "sft_epochs":      "4",
-        # Budget cap: early stopping will fire well before step 500 if the
-        # policy converges (reward_std < 0.05 for 50 consecutive steps).
-        "num_steps":       "500",
+        "num_steps":       "200",
         "num_generations": "4",
         "max_completion":  "384",
         "hub_final_repo":  "Pratyush-01/physix-3b-rl",
         "hub_ckpt_repo":   "Pratyush-01/physix-3b-rl-ckpt",
-        "sft_run_name":    "physix-sft-3b-v4",
-        "grpo_run_name":   "physix-grpo-3b-lr1e5",
+        "sft_run_name":    "physix-sft-3b-final",
+        "grpo_run_name":   "physix-grpo-3b-final",
     },
     "7b": {
         "base_model":      "Qwen/Qwen2.5-7B-Instruct",