Training in progress, step 1000

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1de8090ce06ee283276a4164fb3fc1cbb41f07f3933cb5b30a7a520d950cd8bb
+oid sha256:ab3467fe7b706783d195082717277e7993176e96abb7f168a29914f5a59ec505
 size 268290900

run-1/checkpoint-1000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:866be8cdc735bbe36a639244e641a1ac7cb151100c189abd5e0204a9b16f0c3b
+oid sha256:ab3467fe7b706783d195082717277e7993176e96abb7f168a29914f5a59ec505
 size 268290900

run-1/checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0a5957e7a6bfaac777273b29656197645c0696c422bfdd157dc5bc4f63c4e05a
+oid sha256:055188a1969f89bad7cabd4dce4c1ace96fe3963fb776164fed4776bce4fdc97
 size 536643898

run-1/checkpoint-1000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:60ef01273aaa599804d51a93c6c0c61874ebd50ab9e135a21aab5f7cd0e6487a
+oid sha256:c794bc4c67ef18245dd516031ce405ab557e4d551d225d8dd1e1abc0f2be8e33
 size 1064

run-1/checkpoint-1000/trainer_state.json

@@ -10,50 +10,50 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5980645161290322,
-      "eval_loss": 0.2370469570159912,
-      "eval_runtime": 5.131,
-      "eval_samples_per_second": 604.169,
-      "eval_steps_per_second": 12.668,
+      "eval_accuracy": 0.5838709677419355,
+      "eval_loss": 0.19346614181995392,
+      "eval_runtime": 5.1609,
+      "eval_samples_per_second": 600.675,
+      "eval_steps_per_second": 12.595,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5828151106834412,
-      "learning_rate": 1.4758909853249476e-05,
-      "loss": 0.3737,
+      "grad_norm": 0.5179193019866943,
+      "learning_rate": 1.685534591194969e-05,
+      "loss": 0.3125,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8264516129032258,
-      "eval_loss": 0.11328593641519547,
-      "eval_runtime": 5.5375,
-      "eval_samples_per_second": 559.823,
-      "eval_steps_per_second": 11.738,
+      "eval_accuracy": 0.82,
+      "eval_loss": 0.09356740862131119,
+      "eval_runtime": 5.2878,
+      "eval_samples_per_second": 586.255,
+      "eval_steps_per_second": 12.292,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8780645161290322,
-      "eval_loss": 0.0755920559167862,
-      "eval_runtime": 5.1832,
-      "eval_samples_per_second": 598.091,
-      "eval_steps_per_second": 12.541,
+      "eval_accuracy": 0.8790322580645161,
+      "eval_loss": 0.06245172396302223,
+      "eval_runtime": 5.0672,
+      "eval_samples_per_second": 611.776,
+      "eval_steps_per_second": 12.828,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.5158044099807739,
-      "learning_rate": 9.517819706498952e-06,
-      "loss": 0.1316,
+      "grad_norm": 0.4341842532157898,
+      "learning_rate": 1.371069182389937e-05,
+      "loss": 0.1085,
       "step": 1000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 1908,
+  "max_steps": 3180,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 6,
+  "num_train_epochs": 10,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -71,8 +71,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.4750594581069757,
-    "num_train_epochs": 6,
-    "temperature": 5
+    "alpha": 0.5365842802514643,
+    "num_train_epochs": 10,
+    "temperature": 14
   }
 }

run-1/checkpoint-1000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4b3c3e04f27e1abfa351236657f60e84eba5fb921cf293902231af05b1d43c52
+oid sha256:aff76e4271bd0bf9763ec56f904c0eab2f55fe172f25f099aa8c57c7f7749007
 size 5240

runs/Oct11_17-03-17_c179252869f4/events.out.tfevents.1728667239.c179252869f4.732.2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:820d379076a01b53ffef66bbe4800d9d4c680409363b24b7f8b2944a79b2ed89
-size 13299
+oid sha256:10367c984244420d47ea408506189cbe0d6f5dd6dabfc78adb56bcb4be98ab6e
+size 13833