Model save

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 1.997211509414953e+17,
-    "train_loss": 9.506409932638853,
-    "train_runtime": 976.3941,
+    "train_loss": 9.505996913400315,
+    "train_runtime": 975.3365,
     "train_samples": 46801,
-    "train_samples_per_second": 8.574,
+    "train_samples_per_second": 8.584,
     "train_steps_per_second": 0.134
 }

runs/Nov14_22-02-41_main-ah-train-0-0/events.out.tfevents.1731640323.main-ah-train-0-0.519.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7de6127186cd67945b656dff01a1d9b593b09c6518c7db7ca47a69604b9ffbd
-size 12244
+oid sha256:ec27e36e0174d8b469b4de51608476e6336d06240c28b46174cb93ef8ead573d
+size 12869

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
     "total_flos": 1.997211509414953e+17,
-    "train_loss": 9.506409932638853,
-    "train_runtime": 976.3941,
+    "train_loss": 9.505996913400315,
+    "train_runtime": 975.3365,
     "train_samples": 46801,
-    "train_samples_per_second": 8.574,
+    "train_samples_per_second": 8.584,
     "train_steps_per_second": 0.134
 }

trainer_state.json

@@ -10,208 +10,208 @@
   "log_history": [
     {
       "epoch": 0.007633587786259542,
-      "grad_norm": 177.04702758789062,
+      "grad_norm": 177.04722595214844,
       "learning_rate": 1.4285714285714285e-05,
       "loss": 47.6977,
       "step": 1
     },
     {
       "epoch": 0.03816793893129771,
-      "grad_norm": 85.58233642578125,
+      "grad_norm": 85.62853240966797,
       "learning_rate": 7.142857142857143e-05,
-      "loss": 43.384,
+      "loss": 43.3882,
       "step": 5
     },
     {
       "epoch": 0.07633587786259542,
-      "grad_norm": 22.778974533081055,
+      "grad_norm": 22.000961303710938,
       "learning_rate": 0.00014285714285714287,
-      "loss": 32.4846,
+      "loss": 32.4726,
       "step": 10
     },
     {
       "epoch": 0.11450381679389313,
-      "grad_norm": 12.44094181060791,
+      "grad_norm": 12.63040828704834,
       "learning_rate": 0.00019996395276708856,
-      "loss": 26.3676,
+      "loss": 26.3526,
       "step": 15
     },
     {
       "epoch": 0.15267175572519084,
-      "grad_norm": 6.125822067260742,
+      "grad_norm": 6.1236572265625,
       "learning_rate": 0.00019870502626379127,
-      "loss": 23.7984,
+      "loss": 23.8025,
       "step": 20
     },
     {
       "epoch": 0.19083969465648856,
-      "grad_norm": 9.35146427154541,
+      "grad_norm": 9.36489200592041,
       "learning_rate": 0.00019566964208274254,
-      "loss": 22.3143,
+      "loss": 22.3186,
       "step": 25
     },
     {
       "epoch": 0.22900763358778625,
-      "grad_norm": 16.801265716552734,
+      "grad_norm": 17.107019424438477,
       "learning_rate": 0.0001909124299802724,
-      "loss": 20.7786,
+      "loss": 20.77,
       "step": 30
     },
     {
       "epoch": 0.26717557251908397,
-      "grad_norm": 27.962474822998047,
+      "grad_norm": 27.212358474731445,
       "learning_rate": 0.0001845190085543795,
-      "loss": 17.8884,
+      "loss": 17.8624,
       "step": 35
     },
     {
       "epoch": 0.3053435114503817,
-      "grad_norm": 36.50389099121094,
+      "grad_norm": 36.82498550415039,
       "learning_rate": 0.0001766044443118978,
-      "loss": 12.7762,
+      "loss": 12.7998,
       "step": 40
     },
     {
       "epoch": 0.3435114503816794,
-      "grad_norm": 29.48056983947754,
+      "grad_norm": 30.142446517944336,
       "learning_rate": 0.00016731118074275704,
-      "loss": 7.3657,
+      "loss": 7.393,
       "step": 45
     },
     {
       "epoch": 0.3816793893129771,
-      "grad_norm": 13.53995418548584,
+      "grad_norm": 15.062440872192383,
       "learning_rate": 0.00015680647467311557,
-      "loss": 4.1446,
+      "loss": 4.1547,
       "step": 50
     },
     {
       "epoch": 0.4198473282442748,
-      "grad_norm": 9.89051342010498,
+      "grad_norm": 9.832117080688477,
       "learning_rate": 0.00014527938603696376,
-      "loss": 3.4783,
+      "loss": 3.463,
       "step": 55
     },
     {
       "epoch": 0.4580152671755725,
-      "grad_norm": 4.344300270080566,
+      "grad_norm": 4.392879009246826,
       "learning_rate": 0.00013293737524320797,
-      "loss": 2.8938,
+      "loss": 2.8845,
       "step": 60
     },
     {
       "epoch": 0.4961832061068702,
-      "grad_norm": 1.8051707744598389,
+      "grad_norm": 2.260551929473877,
       "learning_rate": 0.00012000256937760445,
-      "loss": 2.6057,
+      "loss": 2.5922,
       "step": 65
     },
     {
       "epoch": 0.5343511450381679,
-      "grad_norm": 3.4716169834136963,
+      "grad_norm": 3.587684154510498,
       "learning_rate": 0.00010670776443910024,
-      "loss": 2.3928,
+      "loss": 2.3901,
       "step": 70
     },
     {
       "epoch": 0.5725190839694656,
-      "grad_norm": 2.3160088062286377,
+      "grad_norm": 2.6524131298065186,
       "learning_rate": 9.329223556089975e-05,
-      "loss": 2.3088,
+      "loss": 2.3052,
       "step": 75
     },
     {
       "epoch": 0.6106870229007634,
-      "grad_norm": 1.0746971368789673,
+      "grad_norm": 0.9529216885566711,
       "learning_rate": 7.999743062239557e-05,
-      "loss": 2.1992,
+      "loss": 2.2007,
       "step": 80
     },
     {
       "epoch": 0.648854961832061,
-      "grad_norm": 1.997909665107727,
+      "grad_norm": 1.3604825735092163,
       "learning_rate": 6.706262475679205e-05,
-      "loss": 2.152,
+      "loss": 2.1535,
       "step": 85
     },
     {
       "epoch": 0.6870229007633588,
-      "grad_norm": 1.3188209533691406,
+      "grad_norm": 1.0362510681152344,
       "learning_rate": 5.472061396303629e-05,
-      "loss": 2.1218,
+      "loss": 2.1222,
       "step": 90
     },
     {
       "epoch": 0.7251908396946565,
-      "grad_norm": 0.9503675699234009,
+      "grad_norm": 1.1193585395812988,
       "learning_rate": 4.3193525326884435e-05,
-      "loss": 2.0813,
+      "loss": 2.0834,
       "step": 95
     },
     {
       "epoch": 0.7633587786259542,
-      "grad_norm": 1.3065646886825562,
+      "grad_norm": 1.9800268411636353,
       "learning_rate": 3.268881925724297e-05,
-      "loss": 2.0714,
+      "loss": 2.0757,
       "step": 100
     },
     {
       "epoch": 0.8015267175572519,
-      "grad_norm": 1.3419970273971558,
+      "grad_norm": 1.1855801343917847,
       "learning_rate": 2.339555568810221e-05,
-      "loss": 2.0359,
+      "loss": 2.0367,
       "step": 105
     },
     {
       "epoch": 0.8396946564885496,
-      "grad_norm": 2.586273670196533,
+      "grad_norm": 1.8571408987045288,
       "learning_rate": 1.5480991445620542e-05,
       "loss": 2.0068,
       "step": 110
     },
     {
       "epoch": 0.8778625954198473,
-      "grad_norm": 0.914634108543396,
+      "grad_norm": 0.9461548924446106,
       "learning_rate": 9.08757001972762e-06,
-      "loss": 2.0416,
+      "loss": 2.0445,
       "step": 115
     },
     {
       "epoch": 0.916030534351145,
-      "grad_norm": 2.010348081588745,
+      "grad_norm": 2.1497249603271484,
       "learning_rate": 4.3303579172574885e-06,
-      "loss": 2.0116,
+      "loss": 2.0147,
       "step": 120
     },
     {
       "epoch": 0.9541984732824428,
-      "grad_norm": 1.2241454124450684,
+      "grad_norm": 1.3707141876220703,
       "learning_rate": 1.2949737362087156e-06,
-      "loss": 2.032,
+      "loss": 2.0338,
       "step": 125
     },
     {
       "epoch": 0.9923664122137404,
-      "grad_norm": 1.092801809310913,
+      "grad_norm": 1.4225437641143799,
       "learning_rate": 3.60472329114625e-08,
-      "loss": 2.0377,
+      "loss": 2.0402,
       "step": 130
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.032930850982666,
-      "eval_runtime": 23.8518,
-      "eval_samples_per_second": 38.614,
-      "eval_steps_per_second": 1.216,
+      "eval_loss": 2.035097599029541,
+      "eval_runtime": 20.9668,
+      "eval_samples_per_second": 43.927,
+      "eval_steps_per_second": 0.954,
       "step": 131
     },
     {
       "epoch": 1.0,
       "step": 131,
       "total_flos": 1.997211509414953e+17,
-      "train_loss": 9.506409932638853,
-      "train_runtime": 976.3941,
-      "train_samples_per_second": 8.574,
+      "train_loss": 9.505996913400315,
+      "train_runtime": 975.3365,
+      "train_samples_per_second": 8.584,
       "train_steps_per_second": 0.134
     }
   ],