Training in progress, step 500, checkpoint

checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c60f2c33ee44d338900a8f10fa67198f322c3962304a0e6fcd2543d959107fc
+oid sha256:826937018fb6dc7262af840e9f1a98d4522a58bce165fb16fe3aad01437cdbc7
 size 577789320

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76a18dd357e325024a773c7c71745d2bfe864be99fcf3fe22446e730ca917b5f
+oid sha256:a14a7da3480f784949b092c463ef9a05a4eeb1abab99f6aa0dcff986d4624998
 size 1155772233

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e78f22aecb7b127c5adf59cbbbd214d427bb216f6dd5b7bc5fde91b7cbf409e7
+oid sha256:eac4187151331fe6d1ef42bcbe4a8b1d7962c231e4e2a60814d68aa84598d928
 size 14244

checkpoint-500/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.4741882085800171,
+  "best_metric": 0.4463141858577728,
   "best_model_checkpoint": "mikhail-panzo/fil_b32_le3_s4000/checkpoint-500",
-  "epoch": 10.989010989010989,
+  "epoch": 11.11111111111111,
   "eval_steps": 500,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -9,81 +9,81 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.098901098901099,
-      "grad_norm": 2.8254339694976807,
+      "epoch": 1.1111111111111112,
+      "grad_norm": 3.548590898513794,
       "learning_rate": 2.5e-05,
-      "loss": 0.736,
+      "loss": 0.7354,
       "step": 50
     },
     {
-      "epoch": 2.197802197802198,
-      "grad_norm": 4.995375633239746,
+      "epoch": 2.2222222222222223,
+      "grad_norm": 5.517350196838379,
       "learning_rate": 4.9500000000000004e-05,
-      "loss": 0.5913,
+      "loss": 0.5892,
       "step": 100
     },
     {
-      "epoch": 3.2967032967032965,
-      "grad_norm": 2.352879285812378,
+      "epoch": 3.3333333333333335,
+      "grad_norm": 4.143002033233643,
       "learning_rate": 7.45e-05,
-      "loss": 0.5174,
+      "loss": 0.532,
       "step": 150
     },
     {
-      "epoch": 4.395604395604396,
-      "grad_norm": 4.0169596672058105,
+      "epoch": 4.444444444444445,
+      "grad_norm": 6.414477825164795,
       "learning_rate": 9.95e-05,
-      "loss": 0.5275,
+      "loss": 0.5284,
       "step": 200
     },
     {
-      "epoch": 5.4945054945054945,
-      "grad_norm": 5.557097434997559,
+      "epoch": 5.555555555555555,
+      "grad_norm": 2.265143394470215,
       "learning_rate": 0.0001245,
-      "loss": 0.5231,
+      "loss": 0.517,
       "step": 250
     },
     {
-      "epoch": 6.593406593406593,
-      "grad_norm": 8.755484580993652,
+      "epoch": 6.666666666666667,
+      "grad_norm": 4.954311847686768,
       "learning_rate": 0.0001495,
-      "loss": 0.5127,
+      "loss": 0.5295,
       "step": 300
     },
     {
-      "epoch": 7.6923076923076925,
-      "grad_norm": 2.5811164379119873,
+      "epoch": 7.777777777777778,
+      "grad_norm": 2.5993571281433105,
       "learning_rate": 0.00017449999999999999,
-      "loss": 0.5179,
+      "loss": 0.517,
       "step": 350
     },
     {
-      "epoch": 8.791208791208792,
-      "grad_norm": 7.064955234527588,
+      "epoch": 8.88888888888889,
+      "grad_norm": 3.27908992767334,
       "learning_rate": 0.00019950000000000002,
-      "loss": 0.4961,
+      "loss": 0.4939,
       "step": 400
     },
     {
-      "epoch": 9.89010989010989,
-      "grad_norm": 8.559194564819336,
+      "epoch": 10.0,
+      "grad_norm": 4.22811222076416,
       "learning_rate": 0.0002245,
-      "loss": 0.4976,
+      "loss": 0.5043,
       "step": 450
     },
     {
-      "epoch": 10.989010989010989,
-      "grad_norm": 7.322611331939697,
+      "epoch": 11.11111111111111,
+      "grad_norm": 2.7051336765289307,
       "learning_rate": 0.0002495,
-      "loss": 0.502,
+      "loss": 0.5028,
       "step": 500
     },
     {
-      "epoch": 10.989010989010989,
-      "eval_loss": 0.4741882085800171,
-      "eval_runtime": 7.5067,
-      "eval_samples_per_second": 21.448,
-      "eval_steps_per_second": 2.798,
+      "epoch": 11.11111111111111,
+      "eval_loss": 0.4463141858577728,
+      "eval_runtime": 16.8864,
+      "eval_samples_per_second": 9.416,
+      "eval_steps_per_second": 1.184,
       "step": 500
     }
   ],
@@ -104,7 +104,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3684185358544944.0,
+  "total_flos": 3430182513827520.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7108ae4bf8bd54061271f16e95ebc27c1c713cee5a07308d46dd91440b38577c
+oid sha256:86205dd0f5e4c109b754d1a2e96d6be52f2878290f435671b3cc191b44de8f2c
 size 5304