Autonomous Space trainer update

effective_run_config.json

@@ -1,7 +1,7 @@
 {
   "ce_weight_end": 0.5,
   "ce_weight_start": 0.35,
-  "created_at": "2026-04-27T04:16:07.133586+00:00",
+  "created_at": "2026-05-23T06:11:51.769931+00:00",
   "dataset_id": "NorthernTribe-Research/UMSR-v1",
   "distill_enabled": true,
   "enforce_inhouse_models": true,
@@ -25,7 +25,7 @@
   "model_dtype": "bfloat16",
   "native_strict_mode": false,
   "native_trainer_mode": true,
-  "output_dir": "/app/runs/20260427_041552",
+  "output_dir": "/app/runs/20260523_061124",
   "required_bins": [
     "bash",
     "python3",
@@ -51,9 +51,9 @@
       "git": "/usr/bin/git",
       "python3": "/usr/local/bin/python3"
     },
-    "collected_at": "2026-04-27T04:15:59.632937+00:00",
+    "collected_at": "2026-05-23T06:11:39.136659+00:00",
     "cpu_count": 16,
-    "disk_free_gb": 247.15,
+    "disk_free_gb": 326.3,
     "disk_total_gb": 1699.93,
     "in_venv": false,
     "is_ubuntu": false,
@@ -77,7 +77,7 @@
   },
   "save_total_limit": 4,
   "student_model": "NorthernTribe-Research/UMSR-Reasoner-7B",
-  "system_snapshot_path": "/app/runs/20260427_041552/system_snapshot.json",
+  "system_snapshot_path": "/app/runs/20260523_061124/system_snapshot.json",
   "target_repo_id": "NorthernTribe-Research/UMSR-Reasoner-7B",
   "teacher_dtype": "bfloat16",
   "teacher_models": [

live_progress.json

@@ -34,9 +34,9 @@
       "git": "/usr/bin/git",
       "python3": "/usr/local/bin/python3"
     },
-    "collected_at": "2026-04-27T04:15:59.632937+00:00",
+    "collected_at": "2026-05-23T06:11:39.136659+00:00",
     "cpu_count": 16,
-    "disk_free_gb": 247.15,
+    "disk_free_gb": 326.3,
     "disk_total_gb": 1699.93,
     "in_venv": false,
     "is_ubuntu": false,
@@ -59,5 +59,5 @@
     ]
   },
   "status": "completed",
-  "updated_at": "2026-04-27T04:20:14.164886+00:00"
+  "updated_at": "2026-05-23T06:27:43.871485+00:00"
 }

metrics/eval_metrics.json

@@ -1,7 +1,7 @@
 {
   "eval_loss": 5.438441753387451,
-  "eval_runtime": 11.7089,
+  "eval_runtime": 47.7421,
   "eval_samples": 64,
-  "eval_samples_per_second": 5.466,
-  "eval_steps_per_second": 5.466
+  "eval_samples_per_second": 1.341,
+  "eval_steps_per_second": 1.341
 }

metrics/train_metrics.json

@@ -9,8 +9,8 @@
   "temperature_start": 2.5,
   "total_flos": 42322071132.0,
   "train_loss": 4.595640664920211,
-  "train_runtime": 234.9375,
+  "train_runtime": 903.6367,
   "train_samples": 256,
-  "train_samples_per_second": 1.09,
-  "train_steps_per_second": 1.09
+  "train_samples_per_second": 0.283,
+  "train_steps_per_second": 0.283
 }

run_summary.json

@@ -10,13 +10,13 @@
   "distill_enabled": true,
   "enforce_inhouse_models": true,
   "eval_rows": 64,
-  "finished_at": "2026-04-27T04:20:14.164461+00:00",
+  "finished_at": "2026-05-23T06:27:43.870807+00:00",
   "fp16": false,
   "gradient_checkpointing": true,
   "kd_weight_end": 0.5,
   "kd_weight_start": 0.65,
-  "live_events_path": "/app/runs/20260427_041552/live_events.jsonl",
-  "live_progress_path": "/app/runs/20260427_041552/live_progress.json",
+  "live_events_path": "/app/runs/20260523_061124/live_events.jsonl",
+  "live_progress_path": "/app/runs/20260523_061124/live_progress.json",
   "lora_alpha": 64,
   "lora_dropout": 0.05,
   "lora_enabled": true,
@@ -34,7 +34,7 @@
   "mps_available": false,
   "native_strict_mode": false,
   "native_trainer_mode": true,
-  "output_dir": "/app/runs/20260427_041552",
+  "output_dir": "/app/runs/20260523_061124",
   "requested_warmup_steps": 0,
   "required_bins": [
     "bash",
@@ -61,9 +61,9 @@
       "git": "/usr/bin/git",
       "python3": "/usr/local/bin/python3"
     },
-    "collected_at": "2026-04-27T04:15:59.632937+00:00",
+    "collected_at": "2026-05-23T06:11:39.136659+00:00",
     "cpu_count": 16,
-    "disk_free_gb": 247.15,
+    "disk_free_gb": 326.3,
     "disk_total_gb": 1699.93,
     "in_venv": false,
     "is_ubuntu": false,
@@ -86,7 +86,7 @@
     ]
   },
   "save_total_limit": 4,
-  "system_snapshot_path": "/app/runs/20260427_041552/system_snapshot.json",
+  "system_snapshot_path": "/app/runs/20260523_061124/system_snapshot.json",
   "target_repo_id": "NorthernTribe-Research/UMSR-Reasoner-7B",
   "teacher_count": 1,
   "teacher_dtype": "bfloat16",

system_snapshot.json

@@ -6,9 +6,9 @@
     "git": "/usr/bin/git",
     "python3": "/usr/local/bin/python3"
   },
-  "collected_at": "2026-04-27T04:15:59.632937+00:00",
+  "collected_at": "2026-05-23T06:11:39.136659+00:00",
   "cpu_count": 16,
-  "disk_free_gb": 247.15,
+  "disk_free_gb": 326.3,
   "disk_total_gb": 1699.93,
   "in_venv": false,
   "is_ubuntu": false,

trainer_state.json

@@ -317,9 +317,9 @@
       "distill_temperature": 2.373046875,
       "epoch": 0.09765625,
       "eval_loss": 3.927885055541992,
-      "eval_runtime": 13.7367,
-      "eval_samples_per_second": 4.659,
-      "eval_steps_per_second": 4.659,
+      "eval_runtime": 52.4834,
+      "eval_samples_per_second": 1.219,
+      "eval_steps_per_second": 1.219,
       "step": 25
     },
     {
@@ -630,9 +630,9 @@
       "distill_temperature": 2.24609375,
       "epoch": 0.1953125,
       "eval_loss": 4.088868141174316,
-      "eval_runtime": 12.2666,
-      "eval_samples_per_second": 5.217,
-      "eval_steps_per_second": 5.217,
+      "eval_runtime": 42.0977,
+      "eval_samples_per_second": 1.52,
+      "eval_steps_per_second": 1.52,
       "step": 50
     },
     {
@@ -943,9 +943,9 @@
       "distill_temperature": 2.119140625,
       "epoch": 0.29296875,
       "eval_loss": 4.2497992515563965,
-      "eval_runtime": 12.1621,
-      "eval_samples_per_second": 5.262,
-      "eval_steps_per_second": 5.262,
+      "eval_runtime": 44.7873,
+      "eval_samples_per_second": 1.429,
+      "eval_steps_per_second": 1.429,
       "step": 75
     },
     {
@@ -1256,9 +1256,9 @@
       "distill_temperature": 1.9921875,
       "epoch": 0.390625,
       "eval_loss": 4.411334991455078,
-      "eval_runtime": 12.2567,
-      "eval_samples_per_second": 5.222,
-      "eval_steps_per_second": 5.222,
+      "eval_runtime": 43.2827,
+      "eval_samples_per_second": 1.479,
+      "eval_steps_per_second": 1.479,
       "step": 100
     },
     {
@@ -1569,9 +1569,9 @@
       "distill_temperature": 1.865234375,
       "epoch": 0.48828125,
       "eval_loss": 4.573975086212158,
-      "eval_runtime": 12.0166,
-      "eval_samples_per_second": 5.326,
-      "eval_steps_per_second": 5.326,
+      "eval_runtime": 50.3205,
+      "eval_samples_per_second": 1.272,
+      "eval_steps_per_second": 1.272,
       "step": 125
     },
     {
@@ -1882,9 +1882,9 @@
       "distill_temperature": 1.73828125,
       "epoch": 0.5859375,
       "eval_loss": 4.739337921142578,
-      "eval_runtime": 12.1805,
-      "eval_samples_per_second": 5.254,
-      "eval_steps_per_second": 5.254,
+      "eval_runtime": 52.7324,
+      "eval_samples_per_second": 1.214,
+      "eval_steps_per_second": 1.214,
       "step": 150
     },
     {
@@ -2195,9 +2195,9 @@
       "distill_temperature": 1.611328125,
       "epoch": 0.68359375,
       "eval_loss": 4.90593957901001,
-      "eval_runtime": 12.8603,
-      "eval_samples_per_second": 4.977,
-      "eval_steps_per_second": 4.977,
+      "eval_runtime": 44.0685,
+      "eval_samples_per_second": 1.452,
+      "eval_steps_per_second": 1.452,
       "step": 175
     },
     {
@@ -2508,9 +2508,9 @@
       "distill_temperature": 1.484375,
       "epoch": 0.78125,
       "eval_loss": 5.072885513305664,
-      "eval_runtime": 12.48,
-      "eval_samples_per_second": 5.128,
-      "eval_steps_per_second": 5.128,
+      "eval_runtime": 47.8623,
+      "eval_samples_per_second": 1.337,
+      "eval_steps_per_second": 1.337,
       "step": 200
     },
     {
@@ -2821,9 +2821,9 @@
       "distill_temperature": 1.357421875,
       "epoch": 0.87890625,
       "eval_loss": 5.237745761871338,
-      "eval_runtime": 11.6819,
-      "eval_samples_per_second": 5.479,
-      "eval_steps_per_second": 5.479,
+      "eval_runtime": 43.4983,
+      "eval_samples_per_second": 1.471,
+      "eval_steps_per_second": 1.471,
       "step": 225
     },
     {
@@ -3134,9 +3134,9 @@
       "distill_temperature": 1.23046875,
       "epoch": 0.9765625,
       "eval_loss": 5.399942398071289,
-      "eval_runtime": 12.6688,
-      "eval_samples_per_second": 5.052,
-      "eval_steps_per_second": 5.052,
+      "eval_runtime": 46.0879,
+      "eval_samples_per_second": 1.389,
+      "eval_steps_per_second": 1.389,
       "step": 250
     },
     {
@@ -3219,9 +3219,9 @@
       "distill_temperature": 1.2,
       "epoch": 1.0,
       "eval_loss": 5.438441753387451,
-      "eval_runtime": 11.8104,
-      "eval_samples_per_second": 5.419,
-      "eval_steps_per_second": 5.419,
+      "eval_runtime": 49.2057,
+      "eval_samples_per_second": 1.301,
+      "eval_steps_per_second": 1.301,
       "step": 256
     },
     {
@@ -3234,9 +3234,9 @@
       "step": 256,
       "total_flos": 42322071132.0,
       "train_loss": 4.595640664920211,
-      "train_runtime": 234.9375,
-      "train_samples_per_second": 1.09,
-      "train_steps_per_second": 1.09
+      "train_runtime": 903.6367,
+      "train_samples_per_second": 0.283,
+      "train_steps_per_second": 0.283
     }
   ],
   "logging_steps": 1,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9283be6019ca75543608189954ea37f3a9d3c09facf137104cdbb8eb9705f756
+oid sha256:8142320c88a807a5d941541c8191106b80cb97ed3a35e274450030120b9a5ae7
 size 5201