Final adapter for shaer_grpo_20260411_223409

generations_export_summary.json

@@ -1,10 +1,10 @@
 {
   "enabled": true,
   "repo_id": "Shaer-AI/Shaer-adapters-grpo-generations",
-  "run_export_dir": "/root/workspace/Shaer/grpo/outputs/hf_generations_exports/sanity_20260411_174336/runs/sanity_20260411_174336",
-  "chain_export_dir": "/root/workspace/Shaer/grpo/outputs/hf_generations_exports/sanity_20260411_174336/chains/sanity_20260411_174336",
-  "chain_run_count": 1,
-  "run_generation_rows": 32,
-  "run_metric_rows": 5,
-  "chain_metric_rows": 5
+  "run_export_dir": "/root/workspace/Shaer/grpo/outputs/hf_generations_exports/shaer_grpo_20260411_223409/runs/shaer_grpo_20260411_223409",
+  "chain_export_dir": "/root/workspace/Shaer/grpo/outputs/hf_generations_exports/shaer_grpo_20260411_223409/chains/shaer_grpo_20260411_192107",
+  "chain_run_count": 2,
+  "run_generation_rows": 26712,
+  "run_metric_rows": 2704,
+  "chain_metric_rows": 4216
 }

plots/chain_metrics.jsonl

@@ -4212,3 +4212,5 @@
 {"timestamp_utc": "2026-04-12T03:43:26Z", "mode": "train", "global_step": 3298, "epoch": 0.13246575892677834, "loss": 0.0017, "grad_norm": 2.183478593826294, "learning_rate": 9.090909090909092e-09, "num_tokens": 7518824.0, "completions/mean_length": 65.375, "completions/min_length": 65.0, "completions/max_length": 66.0, "completions/clipped_ratio": 0.0, "completions/mean_terminated_length": 65.375, "completions/min_terminated_length": 65.0, "completions/max_terminated_length": 66.0, "rewards/meter/mean": 0.9922501444816589, "rewards/meter/std": 0.004244488663971424, "rewards/count_adherence/mean": 1.0, "rewards/count_adherence/std": 0.0, "rewards/arabic_clean/mean": 1.0, "rewards/arabic_clean/std": 0.0, "rewards/repeat_penalty/mean": 1.0, "rewards/repeat_penalty/std": 0.0, "rewards/total_composite/mean": 0.9922501444816589, "rewards/total_composite/std": 0.004244488663971424, "reward": 0.9922501444816589, "reward_std": 0.0042444937862455845, "frac_reward_zero_std": 0.0, "sampling/sampling_logp_difference/mean": 0.013568882830440998, "sampling/sampling_logp_difference/max": 0.7026901245117188, "sampling/importance_sampling_ratio/min": 0.49525120854377747, "sampling/importance_sampling_ratio/mean": 1.0067819356918335, "sampling/importance_sampling_ratio/max": 1.4506497383117676, "entropy": 0.11512181628495455, "clip_ratio/low_mean": 0.0037878789007663727, "clip_ratio/low_min": 0.0037878789007663727, "clip_ratio/high_mean": 0.005710955825634301, "clip_ratio/high_max": 0.005710955825634301, "clip_ratio/region_mean": 0.009498834726400673, "reward_total_mean": 0.9922501444816589, "reward_meter_mean": 0.9922501444816589, "reward_meter_std": 0.004244488663971424, "reward_count_adherence_mean": 1.0, "reward_count_adherence_std": 0.0, "reward_arabic_clean_mean": 1.0, "reward_arabic_clean_std": 0.0, "reward_repeat_penalty_mean": 1.0, "reward_repeat_penalty_std": 0.0, "reward_total_composite_mean": 0.9922501444816589, "reward_total_composite_std": 0.004244488663971424, "run_id": "shaer_grpo_20260411_223409", "run_sequence_index": 1, "_plot_step": 4131.0}
 {"timestamp_utc": "2026-04-12T03:43:32Z", "mode": "train", "global_step": 3299, "epoch": 0.1325059244085633, "loss": 0.0057, "grad_norm": 3.7170727252960205, "learning_rate": 6.060606060606061e-09, "num_tokens": 7521612.0, "completions/mean_length": 166.5, "completions/min_length": 159.0, "completions/max_length": 170.0, "completions/clipped_ratio": 0.0, "completions/mean_terminated_length": 166.5, "completions/min_terminated_length": 159.0, "completions/max_terminated_length": 170.0, "rewards/meter/mean": 0.9987373352050781, "rewards/meter/std": 0.000354176911059767, "rewards/count_adherence/mean": 1.0, "rewards/count_adherence/std": 0.0, "rewards/arabic_clean/mean": 1.0, "rewards/arabic_clean/std": 0.0, "rewards/repeat_penalty/mean": 1.0, "rewards/repeat_penalty/std": 0.0, "rewards/total_composite/mean": 0.9987373352050781, "rewards/total_composite/std": 0.000354176911059767, "reward": 0.9987373352050781, "reward_std": 0.00035417175968177617, "frac_reward_zero_std": 0.0, "sampling/sampling_logp_difference/mean": 0.03916100785136223, "sampling/sampling_logp_difference/max": 1.9973111152648926, "sampling/importance_sampling_ratio/min": 0.13569968938827515, "sampling/importance_sampling_ratio/mean": 1.001840591430664, "sampling/importance_sampling_ratio/max": 2.0, "entropy": 0.3029062431305647, "clip_ratio/low_mean": 0.01500677247531712, "clip_ratio/low_min": 0.01500677247531712, "clip_ratio/high_mean": 0.015017259865999222, "clip_ratio/high_max": 0.015017259865999222, "clip_ratio/region_mean": 0.030024032341316342, "reward_total_mean": 0.9987373352050781, "reward_meter_mean": 0.9987373352050781, "reward_meter_std": 0.000354176911059767, "reward_count_adherence_mean": 1.0, "reward_count_adherence_std": 0.0, "reward_arabic_clean_mean": 1.0, "reward_arabic_clean_std": 0.0, "reward_repeat_penalty_mean": 1.0, "reward_repeat_penalty_std": 0.0, "reward_total_composite_mean": 0.9987373352050781, "reward_total_composite_std": 0.000354176911059767, "run_id": "shaer_grpo_20260411_223409", "run_sequence_index": 1, "_plot_step": 4132.0}
 {"timestamp_utc": "2026-04-12T03:43:36Z", "mode": "train", "global_step": 3300, "epoch": 0.13254608989034825, "loss": 0.0153, "grad_norm": 6.677674293518066, "learning_rate": 3.0303030303030304e-09, "num_tokens": 7523188.0, "completions/mean_length": 46.0, "completions/min_length": 46.0, "completions/max_length": 46.0, "completions/clipped_ratio": 0.0, "completions/mean_terminated_length": 46.0, "completions/min_terminated_length": 46.0, "completions/max_terminated_length": 46.0, "rewards/meter/mean": 0.9442145824432373, "rewards/meter/std": 0.003297017654404044, "rewards/count_adherence/mean": 1.0, "rewards/count_adherence/std": 0.0, "rewards/arabic_clean/mean": 1.0, "rewards/arabic_clean/std": 0.0, "rewards/repeat_penalty/mean": 1.0, "rewards/repeat_penalty/std": 0.0, "rewards/total_composite/mean": 0.9442145824432373, "rewards/total_composite/std": 0.003297017654404044, "reward": 0.9442145824432373, "reward_std": 0.0032970213796943426, "frac_reward_zero_std": 0.0, "sampling/sampling_logp_difference/mean": 0.021698685362935066, "sampling/sampling_logp_difference/max": 2.1988744735717773, "sampling/importance_sampling_ratio/min": 0.11092793941497803, "sampling/importance_sampling_ratio/mean": 1.0027215480804443, "sampling/importance_sampling_ratio/max": 1.663463830947876, "entropy": 0.09115941543132067, "clip_ratio/low_mean": 0.00815217406488955, "clip_ratio/low_min": 0.00815217406488955, "clip_ratio/high_mean": 0.0027173913549631834, "clip_ratio/high_max": 0.0027173913549631834, "clip_ratio/region_mean": 0.010869565419852734, "reward_total_mean": 0.9442145824432373, "reward_meter_mean": 0.9442145824432373, "reward_meter_std": 0.003297017654404044, "reward_count_adherence_mean": 1.0, "reward_count_adherence_std": 0.0, "reward_arabic_clean_mean": 1.0, "reward_arabic_clean_std": 0.0, "reward_repeat_penalty_mean": 1.0, "reward_repeat_penalty_std": 0.0, "reward_total_composite_mean": 0.9442145824432373, "reward_total_composite_std": 0.003297017654404044, "run_id": "shaer_grpo_20260411_223409", "run_sequence_index": 1, "_plot_step": 4133.0}
+{"timestamp_utc": "2026-04-12T03:45:00Z", "mode": "train", "global_step": 3300, "epoch": 0.13254608989034825, "train_runtime": 18489.879, "train_samples_per_second": 1.428, "train_steps_per_second": 0.178, "total_flos": 0.0, "train_loss": 0.0004697396774788627, "run_id": "shaer_grpo_20260411_223409", "run_sequence_index": 1, "_plot_step": 4133.0}
+{"timestamp_utc": "2026-04-12T03:44:57Z", "mode": "eval", "global_step": 3300, "epoch": 0.13254608989034825, "eval_loss": NaN, "eval_runtime": 80.3156, "eval_samples_per_second": 1.295, "eval_steps_per_second": 0.162, "eval_num_tokens": 7523188.0, "eval_completions/mean_length": 213.14423076923077, "eval_completions/min_length": 61.23076923076923, "eval_completions/max_length": 421.38461538461536, "eval_completions/clipped_ratio": 0.009615384615384616, "eval_completions/mean_terminated_length": 209.98489027756912, "eval_completions/min_terminated_length": 61.23076923076923, "eval_completions/max_terminated_length": 412.2307692307692, "eval_rewards/meter/mean": 0.7804106657321637, "eval_rewards/meter/std": 0.3429693900621854, "eval_rewards/count_adherence/mean": 0.9621203220807589, "eval_rewards/count_adherence/std": 0.06154714152216911, "eval_rewards/arabic_clean/mean": 1.0, "eval_rewards/arabic_clean/std": 0.0, "eval_rewards/repeat_penalty/mean": 0.9574309633328364, "eval_rewards/repeat_penalty/std": 0.07397974411455485, "eval_rewards/total_composite/mean": 0.7239617155148432, "eval_rewards/total_composite/std": 0.3431667788670613, "eval_reward": 0.7239617155148432, "eval_reward_std": NaN, "eval_frac_reward_zero_std": 0.0, "eval_sampling/sampling_logp_difference/mean": 0.03249364231641476, "eval_sampling/sampling_logp_difference/max": 1.2331663278432994, "eval_sampling/importance_sampling_ratio/min": 0.2969068231490942, "eval_sampling/importance_sampling_ratio/mean": 1.0095902956449068, "eval_sampling/importance_sampling_ratio/max": 1.6007587084403405, "eval_entropy": 0.38235602699793303, "eval_clip_ratio/low_mean": 0.0, "eval_clip_ratio/low_min": 0.0, "eval_clip_ratio/high_mean": 0.0, "eval_clip_ratio/high_max": 0.0, "eval_clip_ratio/region_mean": 0.0, "eval_reward_total_mean": 0.7239617155148432, "eval_reward_meter_mean": 0.7804106657321637, "eval_reward_meter_std": 0.3429693900621854, "eval_reward_count_adherence_mean": 0.9621203220807589, "eval_reward_count_adherence_std": 0.06154714152216911, "eval_reward_arabic_clean_mean": 1.0, "eval_reward_arabic_clean_std": 0.0, "eval_reward_repeat_penalty_mean": 0.9574309633328364, "eval_reward_repeat_penalty_std": 0.07397974411455485, "eval_reward_total_composite_mean": 0.7239617155148432, "eval_reward_total_composite_std": 0.3431667788670613, "run_id": "shaer_grpo_20260411_223409", "run_sequence_index": 1, "_plot_step": 4133.0}

plotter.log

@@ -3588,3 +3588,13 @@
 [plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/arabic_gate_chain.png
 [plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/meter_by_meter_run.png
 [plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/meter_by_meter_chain.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/reward_panels_train_run.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/reward_panels_eval_run.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/reward_panels_train_chain.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/reward_panels_eval_chain.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/kl_run.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/kl_chain.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/arabic_gate_run.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/arabic_gate_chain.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/meter_by_meter_run.png
+[plot_live_rewards] updated /root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/plots/meter_by_meter_chain.png

run_summary.json

@@ -1,35 +1,35 @@
 {
-  "timestamp_utc": "2026-04-11T17:48:47Z",
-  "mode": "sanity",
-  "created_at_utc": "2026-04-11T17:46:58Z",
-  "run_id": "sanity_20260411_174336",
-  "run_dir": "/root/workspace/Shaer/grpo/outputs/sanity_check/sanity_20260411_174336",
-  "chain_id": "sanity_20260411_174336",
-  "root_run_id": "sanity_20260411_174336",
-  "parent_run_id": "",
-  "parent_run_dir": "",
-  "run_sequence_index": 0,
+  "timestamp_utc": "2026-04-12T03:45:42Z",
+  "mode": "train",
+  "created_at_utc": "2026-04-11T22:36:28Z",
+  "run_id": "shaer_grpo_20260411_223409",
+  "run_dir": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409",
+  "chain_id": "shaer_grpo_20260411_192107",
+  "root_run_id": "shaer_grpo_20260411_192107",
+  "parent_run_id": "shaer_grpo_20260411_192107",
+  "parent_run_dir": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_192107",
+  "run_sequence_index": 1,
   "output_repo": "Shaer-AI/Shaer-adapters-grpo",
   "resume_decision": {
-    "timestamp_utc": "2026-04-11T17:46:58Z",
-    "resume_mode": "fresh",
-    "requested_spec": "",
-    "result": "fresh",
-    "reason": "explicit_fresh",
-    "local_resume_path": null,
+    "timestamp_utc": "2026-04-11T22:36:28Z",
+    "resume_mode": "from_path",
+    "requested_spec": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_192107/checkpoint-650",
+    "result": "resume",
+    "reason": "explicit_local_path",
+    "local_resume_path": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_192107/checkpoint-650",
     "remote_repo": null,
     "remote_prefix": null,
-    "compatibility": "unchecked",
-    "config_fingerprint": "91c7b4340e889a3a285a9b7b3adb2c6025c061c0e14a1dcaa5a2178c5369cf3b"
+    "compatibility": "fingerprint_mismatch",
+    "config_fingerprint": "dddf85cc86110a84f67b563540177d5e59b5c500695e99fe984dfc39dac36c3f"
   },
-  "best_model_checkpoint": "/root/workspace/Shaer/grpo/outputs/sanity_check/sanity_20260411_174336/checkpoint-2",
-  "global_step": 2,
+  "best_model_checkpoint": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/checkpoint-3100",
+  "global_step": 3300,
   "train_metrics": {
-    "train_runtime": 29.2695,
-    "train_samples_per_second": 0.547,
-    "train_steps_per_second": 0.068,
+    "train_runtime": 18489.879,
+    "train_samples_per_second": 1.428,
+    "train_steps_per_second": 0.178,
     "total_flos": 0.0,
-    "train_loss": 0.12583593279123306
+    "train_loss": 0.0004697396774788627
   },
-  "final_adapter_dir": "/root/workspace/Shaer/grpo/outputs/sanity_check/sanity_20260411_174336/final_adapter"
+  "final_adapter_dir": "/root/workspace/Shaer/grpo/outputs/train/shaer_grpo_20260411_223409/final_adapter"
 }

train.log

@@ -2713,3 +2713,4 @@
 2026-04-12 03:43:36,813 | INFO | train_grpo_train | metrics_logged mode=train step=3300
 2026-04-12 03:44:57,274 | INFO | train_grpo_train | metrics_logged mode=eval step=3300
 2026-04-12 03:45:00,970 | INFO | train_grpo_train | metrics_logged mode=train step=3300
+2026-04-12 03:46:19,803 | INFO | train_grpo_train | pushing final adapter to hub repo=Shaer-AI/Shaer-adapters-grpo

train_stdout.log

@@ -10362,3 +10362,1281 @@ The tokenizer has new PAD/BOS/EOS tokens that differ from the model config and g
 
  ...ots/arabic_gate_chain.png: 100%|██████████| 98.8kB / 98.8kB            
 
  .../meter_by_meter_chain.png: 100%|██████████| 1.06MB / 1.06MB            
 
  ...ts/meter_by_meter_run.png: 100%|██████████|  756kB /  756kB            
+No files have been modified since last commit. Skipping to prevent empty commit.
+
+
+
+
  ..._223409/generations.jsonl:   8%|▊         | 47.5MB /  614MB            
+
+
  ..._223409/generations.jsonl:   8%|▊         | 47.5MB /  614MB            
+
+
  ..._223409/generations.jsonl:  15%|█▍        | 90.6MB /  614MB            
+
+
  ..._223409/generations.jsonl:  22%|██▏       |  136MB /  614MB            
+
+
  ..._223409/generations.jsonl:  29%|██▉       |  177MB /  614MB            
+
+
  ..._223409/generations.jsonl:  46%|████▌     |  284MB /  614MB            
+
+
+
  ..._223409/generations.jsonl:  52%|█████▏    |  321MB /  614MB            
+
+
+
  ..._223409/generations.jsonl:  60%|█████▉    |  366MB /  614MB            
+
+
  ..._223409/generations.jsonl:  67%|██████▋   |  413MB /  614MB            
+
+
  ..._223409/generations.jsonl:  85%|████████▍ |  519MB /  614MB            
+
+
+
  ..._223409/generations.jsonl:  91%|█████████ |  560MB /  614MB            
+
+
+
+
  ...23409/generations.parquet: 100%|█████████▉| 53.5MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|█████████▉|  670kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|█████████▉| 27.1kB / 27.2kB            
+
+
  ..._223409/generations.jsonl: 100%|█████████▉|  613MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|█████████▉| 53.5MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|█████████▉|  670kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|█████████▉| 27.1kB / 27.2kB            
+
+
+
  ..._223409/generations.jsonl: 100%|█████████▉|  613MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|█████████▉| 53.5MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|█████████▉|  670kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|█████████▉| 27.1kB / 27.2kB            
+
+
  ..._223409/generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|██████████| 53.8MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|██████████|  673kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|██████████| 27.2kB / 27.2kB            
+
+
+
  ..._223409/generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|██████████| 53.8MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|██████████|  673kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|██████████| 27.2kB / 27.2kB            
+
+
  ..._223409/generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|██████████| 53.8MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|██████████|  673kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|██████████| 27.2kB / 27.2kB            
+
+
  ..._223409/generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
  ...23409/generations.parquet: 100%|██████████| 53.8MB / 53.8MB            
+
+
+
+
  ...11_223409/metrics.parquet: 100%|██████████|  673kB /  673kB            
+
+
+
+
+
  ...409/reward_curves_run.png: 100%|██████████| 27.2kB / 27.2kB            
+
+
  ..._223409/generations.jsonl: 100%|██████████|  614MB /  614MB            
+
  ...23409/generations.parquet: 100%|██████████| 53.8MB / 53.8MB            
+
  ...11_223409/metrics.parquet: 100%|██████████|  673kB /  673kB            
+
  ...409/reward_curves_run.png: 100%|██████████| 27.2kB / 27.2kB            
+
+
+
+
  ...107/chain_metrics.parquet: 100%|██████████|  982kB /  982kB            
+
+
+
  ...7/chain_reward_curves.png: 100%|██████████| 27.8kB / 27.8kB            
+
+
  ...107/chain_metrics.parquet: 100%|██████████|  982kB /  982kB            
+
+
+
  ...7/chain_reward_curves.png: 100%|██████████| 27.8kB / 27.8kB            
+
+
+
  ...107/chain_metrics.parquet: 100%|██████████|  982kB /  982kB            
+
+
+
  ...7/chain_reward_curves.png: 100%|██████████| 27.8kB / 27.8kB            
+
+
  ...107/chain_metrics.parquet: 100%|██████████|  982kB /  982kB            
+
+
+
  ...7/chain_reward_curves.png: 100%|██████████| 27.8kB / 27.8kB            
+
+
  ...107/chain_metrics.parquet: 100%|██████████|  982kB /  982kB            
+
  ...7/chain_reward_curves.png: 100%|██████████| 27.8kB / 27.8kB            
+2026-04-12 03:46:19,803 | INFO | train_grpo_train | pushing final adapter to hub repo=Shaer-AI/Shaer-adapters-grpo
+
+
+
+
  ..._223409/training_args.bin: 100%|██████████| 7.50kB / 7.50kB            
+
+
+
  ...plots/arabic_gate_run.png: 100%|██████████| 81.1kB / 81.1kB            
+
+
+
+
  ...ots/arabic_gate_chain.png: 100%|██████████| 98.8kB / 98.8kB            
+
+
+
+
+
  ...l_adapter/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
  ...adapter/training_args.bin: 100%|██████████| 7.50kB / 7.50kB            
+
+
+
+
+
+
+
  ...1_223409/plots/kl_run.png: 100%|██████████| 24.6kB / 24.6kB            
+
+
+
+
+
+
+
+
  ...223409/plots/kl_chain.png: 100%|██████████| 33.4kB / 33.4kB            
+
+
+
+
+
+
+
+
+
  .../meter_by_meter_chain.png: 100%|██████████| 1.06MB / 1.06MB            
+
+
+
+
+
+
+
+
+
+
  ...rd_panels_train_chain.png: 100%|██████████|  350kB /  350kB            
+
+
+
+
+
+
+
+
+
+
+
  ...adapter_model.safetensors:  10%|█         | 64.0MB /  640MB            
+
+
  ..._223409/training_args.bin: 100%|██████████| 7.50kB / 7.50kB            
+
+
+
  ...plots/arabic_gate_run.png: 100%|██████████| 81.1kB / 81.1kB            
+
+
+
+
  ...ots/arabic_gate_chain.png: 100%|██████████| 98.8kB / 98.8kB            
+
+
+
+
+
  ...l_adapter/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
  ...adapter/training_args.bin: 100%|██████████| 7.50kB / 7.50kB            
+
+
+
+
+
+
+
  ...1_223409/plots/kl_run.png: 100%|██████████| 24.6kB / 24.6kB            
+
+
+
+
+
+
+
+
  ...223409/plots/kl_chain.png: 100%|██████████| 33.4kB / 33.4kB            
+
+
+
+
+
+
+
+
+
  .../meter_by_meter_chain.png: 100%|██████████| 1.06MB / 1.06MB            
+
+
+
+
+
+
+
+
+
+
  ...rd_panels_train_chain.png: 100%|██████████|  350kB /  350kB            
+
+
+
+
+
+
+
+
+
+
+
  [+ 8 files]                 :  18%|█▊        |  440MB / 2.42GB            
+
+
  ...rd_panels_train_chain.png: 100%|██████████|  350kB /  350kB            
+
+
+
  ...adapter_model.safetensors:  20%|██        |  128MB /  640MB            
+
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
+
  ...adapter_model.safetensors:  19%|█▉        |  120MB /  640MB            
+
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
  ...409/all_generations.jsonl:  20%|█▉        |  120MB /  614MB            
+
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  [+ 1 files]                 : 100%|██████████|  107MB /  107MB            
+
+
  ...adapter_model.safetensors:  28%|██▊       |  176MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  28%|██▊       |  176MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  29%|██▊       |  176MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  38%|███▊      |  240MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  38%|███▊      |  240MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  38%|███▊      |  232MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  46%|████▋     |  296MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  46%|████▋     |  296MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  47%|████▋     |  288MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  55%|█████▌    |  352MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  55%|█████▌    |  352MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  56%|█████▌    |  344MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  64%|██████▍   |  408MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  64%|██████▍   |  408MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  65%|██████▌   |  400MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  73%|███████▎  |  464MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  74%|███████▍  |  472MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  74%|███████▍  |  456MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  83%|████████▎ |  528MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  83%|████████▎ |  528MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  85%|████████▍ |  520MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors:  91%|█████████▏|  584MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors:  91%|█████████▏|  584MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl:  94%|█████████▍|  576MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors: 100%|██████████|  640MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors: 100%|██████████|  640MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
  ...adapter_model.safetensors: 100%|██████████|  640MB /  640MB            
+
+
+
  ...ward_panels_train_run.png: 100%|██████████|  342kB /  342kB            
+
+
+
+
  ...adapter_model.safetensors: 100%|██████████|  640MB /  640MB            
+
+
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
+
+
+
  [+ 3 files]                 : 100%|██████████| 2.61MB / 2.61MB            
+
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
+
+
+
  ...ts/meter_by_meter_run.png: 100%|██████████|  754kB /  754kB            
+
+
+
+
+
+
+
+
+
+
  ...ard_panels_eval_chain.png: 100%|██████████|  318kB /  318kB            
+
+
+
+
+
+
+
+
+
+
+
  ...eward_panels_eval_run.png: 100%|██████████|  310kB /  310kB            
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
+
+
+
  ...ts/meter_by_meter_run.png: 100%|██████████|  754kB /  754kB            
+
+
+
+
+
+
+
+
+
+
  ...ard_panels_eval_chain.png: 100%|██████████|  318kB /  318kB            
+
+
+
+
+
+
+
+
+
+
+
  ...eward_panels_eval_run.png: 100%|██████████|  310kB /  310kB            
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
+
+
+
  ...ts/meter_by_meter_run.png: 100%|██████████|  754kB /  754kB            
+
+
+
+
+
+
+
+
+
+
  ...ard_panels_eval_chain.png: 100%|██████████|  318kB /  318kB            
+
+
+
+
+
+
+
+
+
+
+
  ...eward_panels_eval_run.png: 100%|██████████|  310kB /  310kB            
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
+
+
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
+
+
+
+
+
+
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
+
+
+
+
+
+
+
+
  ...ts/meter_by_meter_run.png: 100%|██████████|  754kB /  754kB            
+
+
+
+
+
+
+
+
+
+
  ...ard_panels_eval_chain.png: 100%|██████████|  318kB /  318kB            
+
+
+
+
+
+
+
+
+
+
+
  ...eward_panels_eval_run.png: 100%|██████████|  310kB /  310kB            
+
+
  ...unt_adherence_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
  ...409/all_generations.jsonl: 100%|██████████|  614MB /  614MB            
+
  ..._arabic_clean_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
  .../reward_meter_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
  ...tal_composite_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
  ...epeat_penalty_debug.jsonl: 100%|██████████|  106MB /  106MB            
+
  ...11_223409/tokenizer.model: 100%|██████████| 1.23MB / 1.23MB            
+
  ...ts/meter_by_meter_run.png: 100%|██████████|  754kB /  754kB            
+
  ...ard_panels_eval_chain.png: 100%|██████████|  318kB /  318kB            
+
  ...eward_panels_eval_run.png: 100%|██████████|  310kB /  310kB