Training in progress, step 300, checkpoint

checkpoint-300/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fc3ad9debeaf7f538b711568decb92f96c55a1f9a1ef95eb34dad71e457c13a
+oid sha256:62a9d6a909e960ee80782469ede5ba8d3e2eea7bf30074590bc10c585494e086
 size 9873829

checkpoint-300/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:16fec27a80bcbb1a6e67e293ccf00bca3ebde2e43d755de25979f2fa54de8ded
+oid sha256:24a063ed1093720634af2dc8cfc1f9624b8ea87073f90b903ee23bf495cb2371
 size 42788

checkpoint-300/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:08e4587196cd89ca6e5903b8eba5be6994a08c795e0d94e5441e481200234fe6
+oid sha256:ec9ccd98c036b5a03b30318ebad776b3e96070d234f0645aa1eb655af4cc58dd
 size 14244

checkpoint-300/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4a19207b90f493015dafeb646311768ce48791e750726b118b7593521d25ff42
+oid sha256:c7e2d324ea44b9a89e599afab7280bd8bacd74de84946fdad19e15e3bbb9abf3
 size 1064

checkpoint-300/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.5592061877250671,
+  "best_metric": 0.6528732180595398,
   "best_model_checkpoint": "output/checkpoint-200",
-  "epoch": 0.3026481715006305,
+  "epoch": 0.24064975433670913,
   "eval_steps": 50,
   "global_step": 300,
   "is_hyper_param_search": false,
@@ -10,106 +10,106 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 1.6666666666666667e-06,
-      "loss": 0.5465,
+      "learning_rate": 1.3333333333333332e-06,
+      "loss": 0.785,
       "step": 1
     },
     {
-      "epoch": 0.05,
-      "learning_rate": 7.5e-05,
-      "loss": 0.7053,
+      "epoch": 0.04,
+      "learning_rate": 5.9999999999999995e-05,
+      "loss": 0.6547,
       "step": 50
     },
     {
-      "epoch": 0.05,
-      "eval_accuracy": 0.6561679790026247,
-      "eval_loss": 0.6764907836914062,
-      "eval_runtime": 59.7766,
-      "eval_samples_per_second": 6.374,
-      "eval_steps_per_second": 1.606,
+      "epoch": 0.04,
+      "eval_accuracy": 0.6626865671641791,
+      "eval_loss": 0.8300915956497192,
+      "eval_runtime": 61.5689,
+      "eval_samples_per_second": 5.441,
+      "eval_steps_per_second": 1.364,
       "step": 50
     },
     {
-      "epoch": 0.1,
-      "learning_rate": 0.00015666666666666666,
-      "loss": 0.6576,
+      "epoch": 0.08,
+      "learning_rate": 0.00012666666666666666,
+      "loss": 0.6654,
       "step": 100
     },
     {
-      "epoch": 0.1,
-      "eval_accuracy": 0.7217847769028871,
-      "eval_loss": 0.574826717376709,
-      "eval_runtime": 59.8708,
-      "eval_samples_per_second": 6.364,
-      "eval_steps_per_second": 1.603,
+      "epoch": 0.08,
+      "eval_accuracy": 0.6895522388059702,
+      "eval_loss": 0.6663276553153992,
+      "eval_runtime": 56.4633,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 100
     },
     {
-      "epoch": 0.15,
-      "learning_rate": 0.00023999999999999998,
-      "loss": 0.5874,
+      "epoch": 0.12,
+      "learning_rate": 0.00019333333333333333,
+      "loss": 0.6491,
       "step": 150
     },
     {
-      "epoch": 0.15,
-      "eval_accuracy": 0.7480314960629921,
-      "eval_loss": 0.5501317381858826,
-      "eval_runtime": 55.7487,
-      "eval_samples_per_second": 6.834,
-      "eval_steps_per_second": 1.722,
+      "epoch": 0.12,
+      "eval_accuracy": 0.6537313432835821,
+      "eval_loss": 0.9172552824020386,
+      "eval_runtime": 56.4705,
+      "eval_samples_per_second": 5.932,
+      "eval_steps_per_second": 1.488,
       "step": 150
     },
     {
-      "epoch": 0.2,
-      "learning_rate": 0.0002948780487804878,
-      "loss": 0.5775,
+      "epoch": 0.16,
+      "learning_rate": 0.00026,
+      "loss": 0.641,
       "step": 200
     },
     {
-      "epoch": 0.2,
-      "eval_accuracy": 0.7401574803149606,
-      "eval_loss": 0.5592061877250671,
-      "eval_runtime": 54.7265,
-      "eval_samples_per_second": 6.962,
-      "eval_steps_per_second": 1.754,
+      "epoch": 0.16,
+      "eval_accuracy": 0.6865671641791045,
+      "eval_loss": 0.6528732180595398,
+      "eval_runtime": 56.4485,
+      "eval_samples_per_second": 5.935,
+      "eval_steps_per_second": 1.488,
       "step": 200
     },
     {
-      "epoch": 0.25,
-      "learning_rate": 0.00027658536585365855,
-      "loss": 0.5821,
+      "epoch": 0.2,
+      "learning_rate": 0.0002941234084231146,
+      "loss": 0.6521,
       "step": 250
     },
     {
-      "epoch": 0.25,
-      "eval_accuracy": 0.7664041994750657,
-      "eval_loss": 0.5005777478218079,
-      "eval_runtime": 54.7418,
-      "eval_samples_per_second": 6.96,
-      "eval_steps_per_second": 1.754,
+      "epoch": 0.2,
+      "eval_accuracy": 0.6835820895522388,
+      "eval_loss": 0.8197052478790283,
+      "eval_runtime": 56.4577,
+      "eval_samples_per_second": 5.934,
+      "eval_steps_per_second": 1.488,
       "step": 250
     },
     {
-      "epoch": 0.3,
-      "learning_rate": 0.00025829268292682926,
-      "loss": 0.5699,
+      "epoch": 0.24,
+      "learning_rate": 0.00027943192948090106,
+      "loss": 0.638,
       "step": 300
     },
     {
-      "epoch": 0.3,
-      "eval_accuracy": 0.6745406824146981,
-      "eval_loss": 0.7000961303710938,
-      "eval_runtime": 54.7375,
-      "eval_samples_per_second": 6.96,
-      "eval_steps_per_second": 1.754,
+      "epoch": 0.24,
+      "eval_accuracy": 0.6835820895522388,
+      "eval_loss": 0.6961421370506287,
+      "eval_runtime": 56.4506,
+      "eval_samples_per_second": 5.934,
+      "eval_steps_per_second": 1.488,
       "step": 300
     }
   ],
   "logging_steps": 50,
-  "max_steps": 1000,
-  "num_train_epochs": 2,
+  "max_steps": 1246,
+  "num_train_epochs": 1,
   "save_steps": 100,
-  "total_flos": 5.170853853978624e+16,
+  "total_flos": 1.1215600847192064e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-300/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:303b684398cbe6b39e16627081218289c5098690bdfd2926fac0c4c4b61ed03d
+oid sha256:e32e120c6d65576e6fe76fdf63d907877e27469586cc746d8cd6a71d07635ea9
 size 4472