Training in progress, step 500, checkpoint

checkpoint-500/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc84f9ee33f5890636c001059ed8024eba9f7b951b63ab62fd2202c9d1e25ffb
+oid sha256:266d1d8179f7989f9b36af575e9b402ceb11422222b40cebe99e87903c573107
 size 9873829

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b56684755656fa2ff1211b7c22bf6db2a30451389b81f4be25ec647702d5028
+oid sha256:58865709f250f4ff0501b31b43475fc1203984456d814556216d59d6f4141bde
 size 42788

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:75f760f829b44edae2a2be950380f961c72b52b6d2def775e5da9330c9837056
+oid sha256:5f7653759f4f1ddd9069725e89aaebdc88edf3a28844142a281b371fb055da08
 size 14244

checkpoint-500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4b9105f9ddc416de984dcfbba85753bcd17de130f398fcf8c52b40a5816c0c95
+oid sha256:529f7b5a081c4e5baab9943dcc97123ba1af7e2a5a8a998441f676f9817f2b58
 size 1064

checkpoint-500/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.5232318639755249,
-  "best_model_checkpoint": "output/checkpoint-500",
-  "epoch": 0.5044136191677175,
+  "best_metric": 0.6528732180595398,
+  "best_model_checkpoint": "output/checkpoint-200",
+  "epoch": 0.40108292389451516,
   "eval_steps": 50,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -10,166 +10,166 @@
   "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
     },
     {
-      "epoch": 0.35,
-      "learning_rate": 0.00023999999999999998,
-      "loss": 0.5079,
+      "epoch": 0.28,
+      "learning_rate": 0.00026474045053868754,
+      "loss": 0.6857,
       "step": 350
     },
     {
-      "epoch": 0.35,
-      "eval_accuracy": 0.7322834645669292,
-      "eval_loss": 0.5262107253074646,
-      "eval_runtime": 54.7143,
-      "eval_samples_per_second": 6.963,
-      "eval_steps_per_second": 1.755,
+      "epoch": 0.28,
+      "eval_accuracy": 0.6835820895522388,
+      "eval_loss": 0.7876304388046265,
+      "eval_runtime": 56.4651,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 350
     },
     {
-      "epoch": 0.4,
-      "learning_rate": 0.00022170731707317072,
-      "loss": 0.5826,
+      "epoch": 0.32,
+      "learning_rate": 0.000250048971596474,
+      "loss": 0.6469,
       "step": 400
     },
     {
-      "epoch": 0.4,
-      "eval_accuracy": 0.7296587926509186,
-      "eval_loss": 0.5798280835151672,
-      "eval_runtime": 54.7163,
-      "eval_samples_per_second": 6.963,
-      "eval_steps_per_second": 1.755,
+      "epoch": 0.32,
+      "eval_accuracy": 0.6626865671641791,
+      "eval_loss": 0.7487305402755737,
+      "eval_runtime": 56.4489,
+      "eval_samples_per_second": 5.935,
+      "eval_steps_per_second": 1.488,
       "step": 400
     },
     {
-      "epoch": 0.45,
-      "learning_rate": 0.00020341463414634146,
-      "loss": 0.5009,
+      "epoch": 0.36,
+      "learning_rate": 0.00023535749265426052,
+      "loss": 0.6813,
       "step": 450
     },
     {
-      "epoch": 0.45,
-      "eval_accuracy": 0.7506561679790026,
-      "eval_loss": 0.5315915942192078,
-      "eval_runtime": 54.7235,
-      "eval_samples_per_second": 6.962,
-      "eval_steps_per_second": 1.754,
+      "epoch": 0.36,
+      "eval_accuracy": 0.6626865671641791,
+      "eval_loss": 0.8447557091712952,
+      "eval_runtime": 56.4428,
+      "eval_samples_per_second": 5.935,
+      "eval_steps_per_second": 1.488,
       "step": 450
     },
     {
-      "epoch": 0.5,
-      "learning_rate": 0.00018512195121951218,
-      "loss": 0.4756,
+      "epoch": 0.4,
+      "learning_rate": 0.000220666013712047,
+      "loss": 0.645,
       "step": 500
     },
     {
-      "epoch": 0.5,
-      "eval_accuracy": 0.7349081364829396,
-      "eval_loss": 0.5232318639755249,
-      "eval_runtime": 54.7185,
-      "eval_samples_per_second": 6.963,
-      "eval_steps_per_second": 1.754,
+      "epoch": 0.4,
+      "eval_accuracy": 0.6835820895522388,
+      "eval_loss": 0.6624641418457031,
+      "eval_runtime": 56.4531,
+      "eval_samples_per_second": 5.934,
+      "eval_steps_per_second": 1.488,
       "step": 500
     }
   ],
   "logging_steps": 50,
-  "max_steps": 1000,
-  "num_train_epochs": 2,
+  "max_steps": 1246,
+  "num_train_epochs": 1,
   "save_steps": 100,
-  "total_flos": 8.623199015841792e+16,
+  "total_flos": 1.86852300708864e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

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