Training in progress, step 1200, checkpoint

checkpoint-1200/adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f3c6ad0e1483eeb374e6bad78a21b868c25d951e108d8e2dced40ad5ae81eff
+oid sha256:fb6b29aeb548986b562089e72cbf92656247531576e35b750cbfb58b0b9db622
 size 9873829

checkpoint-1200/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c98c717e48e81c49380bc567fb5dd3dfee9c5d1b68c37d64f9753e2a99ac927e
+oid sha256:a2cd3e6580ac2a5184d141574b0a8c47d8aeda65a979d6c3ccd1df9b23f57565
 size 42788

checkpoint-1200/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:82adbf581ed48a89b7b4c8794c145d9981de136a23bc37ac8a99d5578392dd61
+oid sha256:cd12f4b261be99d546ae5b38a971f17f6631d591c1c703890c4622802d5a7a33
 size 14244

checkpoint-1200/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:482f6657b9b139b6e60d7775735e87c6ebfc5f6d602073a1362c3bdbdd1360ff
+oid sha256:8ac0cc14621ab6f7bb4414c0773b49fa9aadac42a638a51c7570ec726b06f46f
 size 1064

checkpoint-1200/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.6053406596183777,
-  "best_model_checkpoint": "output/checkpoint-500",
-  "epoch": 0.9622130901072466,
+  "best_metric": 0.6430063247680664,
+  "best_model_checkpoint": "output/checkpoint-50",
+  "epoch": 0.9620202425092694,
   "eval_steps": 50,
   "global_step": 1200,
   "is_hyper_param_search": false,
@@ -10,376 +10,376 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 1.3333333333333332e-06,
-      "loss": 0.6133,
+      "learning_rate": 4.444444444444445e-07,
+      "loss": 0.5542,
       "step": 1
     },
     {
       "epoch": 0.04,
-      "learning_rate": 5.9999999999999995e-05,
-      "loss": 0.61,
+      "learning_rate": 2e-05,
+      "loss": 0.554,
       "step": 50
     },
     {
       "epoch": 0.04,
-      "eval_accuracy": 0.6895522388059702,
-      "eval_loss": 0.6718239784240723,
-      "eval_runtime": 61.4305,
-      "eval_samples_per_second": 5.453,
-      "eval_steps_per_second": 1.367,
+      "eval_accuracy": 0.6626865671641791,
+      "eval_loss": 0.6430063247680664,
+      "eval_runtime": 61.7692,
+      "eval_samples_per_second": 5.423,
+      "eval_steps_per_second": 1.36,
       "step": 50
     },
     {
       "epoch": 0.08,
-      "learning_rate": 0.00012666666666666666,
-      "loss": 0.6233,
+      "learning_rate": 4.222222222222222e-05,
+      "loss": 0.5957,
       "step": 100
     },
     {
       "epoch": 0.08,
-      "eval_accuracy": 0.6835820895522388,
-      "eval_loss": 0.695776641368866,
-      "eval_runtime": 61.5801,
-      "eval_samples_per_second": 5.44,
-      "eval_steps_per_second": 1.364,
+      "eval_accuracy": 0.6686567164179105,
+      "eval_loss": 0.6705919504165649,
+      "eval_runtime": 60.4388,
+      "eval_samples_per_second": 5.543,
+      "eval_steps_per_second": 1.39,
       "step": 100
     },
     {
       "epoch": 0.12,
-      "learning_rate": 0.00019333333333333333,
-      "loss": 0.6187,
+      "learning_rate": 6.444444444444446e-05,
+      "loss": 0.5668,
       "step": 150
     },
     {
       "epoch": 0.12,
-      "eval_accuracy": 0.6895522388059702,
-      "eval_loss": 0.6547008156776428,
-      "eval_runtime": 61.4876,
-      "eval_samples_per_second": 5.448,
-      "eval_steps_per_second": 1.366,
+      "eval_accuracy": 0.6805970149253732,
+      "eval_loss": 0.7224913835525513,
+      "eval_runtime": 56.4882,
+      "eval_samples_per_second": 5.93,
+      "eval_steps_per_second": 1.487,
       "step": 150
     },
     {
       "epoch": 0.16,
-      "learning_rate": 0.00026,
-      "loss": 0.667,
+      "learning_rate": 8.666666666666667e-05,
+      "loss": 0.5626,
       "step": 200
     },
     {
       "epoch": 0.16,
-      "eval_accuracy": 0.6686567164179105,
-      "eval_loss": 0.7339211702346802,
-      "eval_runtime": 61.0301,
-      "eval_samples_per_second": 5.489,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.6746268656716418,
+      "eval_loss": 0.7115103006362915,
+      "eval_runtime": 56.506,
+      "eval_samples_per_second": 5.929,
+      "eval_steps_per_second": 1.487,
       "step": 200
     },
     {
       "epoch": 0.2,
-      "learning_rate": 0.0002941291585127201,
-      "loss": 0.6896,
+      "learning_rate": 9.80430528375734e-05,
+      "loss": 0.5604,
       "step": 250
     },
     {
       "epoch": 0.2,
-      "eval_accuracy": 0.6388059701492538,
-      "eval_loss": 0.6455020904541016,
-      "eval_runtime": 61.0356,
-      "eval_samples_per_second": 5.489,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.6746268656716418,
+      "eval_loss": 0.700947105884552,
+      "eval_runtime": 56.4927,
+      "eval_samples_per_second": 5.93,
+      "eval_steps_per_second": 1.487,
       "step": 250
     },
     {
       "epoch": 0.24,
-      "learning_rate": 0.0002794520547945205,
-      "loss": 0.6732,
+      "learning_rate": 9.315068493150684e-05,
+      "loss": 0.5675,
       "step": 300
     },
     {
       "epoch": 0.24,
-      "eval_accuracy": 0.6388059701492538,
-      "eval_loss": 0.8866483569145203,
-      "eval_runtime": 61.5006,
-      "eval_samples_per_second": 5.447,
-      "eval_steps_per_second": 1.366,
+      "eval_accuracy": 0.6597014925373135,
+      "eval_loss": 0.662155032157898,
+      "eval_runtime": 56.4903,
+      "eval_samples_per_second": 5.93,
+      "eval_steps_per_second": 1.487,
       "step": 300
     },
     {
       "epoch": 0.28,
-      "learning_rate": 0.0002647749510763209,
-      "loss": 0.6255,
+      "learning_rate": 8.825831702544032e-05,
+      "loss": 0.5725,
       "step": 350
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.6746268656716418,
-      "eval_loss": 0.6250954866409302,
-      "eval_runtime": 61.4499,
-      "eval_samples_per_second": 5.452,
-      "eval_steps_per_second": 1.367,
+      "eval_accuracy": 0.6567164179104478,
+      "eval_loss": 0.6746364831924438,
+      "eval_runtime": 56.482,
+      "eval_samples_per_second": 5.931,
+      "eval_steps_per_second": 1.487,
       "step": 350
     },
     {
       "epoch": 0.32,
-      "learning_rate": 0.0002500978473581213,
-      "loss": 0.6522,
+      "learning_rate": 8.336594911937378e-05,
+      "loss": 0.5884,
       "step": 400
     },
     {
       "epoch": 0.32,
-      "eval_accuracy": 0.6925373134328359,
-      "eval_loss": 0.7217190861701965,
-      "eval_runtime": 61.1569,
-      "eval_samples_per_second": 5.478,
-      "eval_steps_per_second": 1.374,
+      "eval_accuracy": 0.6656716417910448,
+      "eval_loss": 0.7047849893569946,
+      "eval_runtime": 56.4873,
+      "eval_samples_per_second": 5.931,
+      "eval_steps_per_second": 1.487,
       "step": 400
     },
     {
       "epoch": 0.36,
-      "learning_rate": 0.00023542074363992171,
-      "loss": 0.6592,
+      "learning_rate": 7.847358121330725e-05,
+      "loss": 0.5567,
       "step": 450
     },
     {
       "epoch": 0.36,
-      "eval_accuracy": 0.6417910447761194,
-      "eval_loss": 0.636791467666626,
-      "eval_runtime": 61.431,
-      "eval_samples_per_second": 5.453,
-      "eval_steps_per_second": 1.367,
+      "eval_accuracy": 0.6477611940298508,
+      "eval_loss": 0.6832321286201477,
+      "eval_runtime": 56.4814,
+      "eval_samples_per_second": 5.931,
+      "eval_steps_per_second": 1.487,
       "step": 450
     },
     {
       "epoch": 0.4,
-      "learning_rate": 0.00022074363992172208,
-      "loss": 0.624,
+      "learning_rate": 7.35812133072407e-05,
+      "loss": 0.5723,
       "step": 500
     },
     {
       "epoch": 0.4,
-      "eval_accuracy": 0.7044776119402985,
-      "eval_loss": 0.6053406596183777,
-      "eval_runtime": 61.4573,
-      "eval_samples_per_second": 5.451,
-      "eval_steps_per_second": 1.367,
+      "eval_accuracy": 0.6955223880597015,
+      "eval_loss": 0.7437570691108704,
+      "eval_runtime": 56.4677,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 500
     },
     {
       "epoch": 0.44,
-      "learning_rate": 0.0002060665362035225,
-      "loss": 0.6098,
+      "learning_rate": 6.868884540117417e-05,
+      "loss": 0.5908,
       "step": 550
     },
     {
       "epoch": 0.44,
-      "eval_accuracy": 0.6477611940298508,
-      "eval_loss": 0.8857873678207397,
-      "eval_runtime": 61.4645,
-      "eval_samples_per_second": 5.45,
-      "eval_steps_per_second": 1.367,
+      "eval_accuracy": 0.6238805970149254,
+      "eval_loss": 0.6612433791160583,
+      "eval_runtime": 56.4799,
+      "eval_samples_per_second": 5.931,
+      "eval_steps_per_second": 1.487,
       "step": 550
     },
     {
       "epoch": 0.48,
-      "learning_rate": 0.00019138943248532287,
-      "loss": 0.6357,
+      "learning_rate": 6.379647749510763e-05,
+      "loss": 0.5681,
       "step": 600
     },
     {
       "epoch": 0.48,
-      "eval_accuracy": 0.7074626865671642,
-      "eval_loss": 0.607388436794281,
-      "eval_runtime": 61.4941,
-      "eval_samples_per_second": 5.448,
-      "eval_steps_per_second": 1.366,
+      "eval_accuracy": 0.6507462686567164,
+      "eval_loss": 0.6833732724189758,
+      "eval_runtime": 56.4689,
+      "eval_samples_per_second": 5.932,
+      "eval_steps_per_second": 1.488,
       "step": 600
     },
     {
       "epoch": 0.52,
-      "learning_rate": 0.0001767123287671233,
-      "loss": 0.6103,
+      "learning_rate": 5.89041095890411e-05,
+      "loss": 0.5782,
       "step": 650
     },
     {
       "epoch": 0.52,
-      "eval_accuracy": 0.6746268656716418,
-      "eval_loss": 0.6098934412002563,
-      "eval_runtime": 56.4929,
-      "eval_samples_per_second": 5.93,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.6507462686567164,
+      "eval_loss": 0.67350834608078,
+      "eval_runtime": 56.4637,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 650
     },
     {
       "epoch": 0.56,
-      "learning_rate": 0.00016203522504892366,
-      "loss": 0.6025,
+      "learning_rate": 5.401174168297456e-05,
+      "loss": 0.569,
       "step": 700
     },
     {
       "epoch": 0.56,
-      "eval_accuracy": 0.6895522388059702,
-      "eval_loss": 0.6117912530899048,
-      "eval_runtime": 61.1994,
-      "eval_samples_per_second": 5.474,
-      "eval_steps_per_second": 1.373,
+      "eval_accuracy": 0.6686567164179105,
+      "eval_loss": 0.7085319757461548,
+      "eval_runtime": 56.4627,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 700
     },
     {
       "epoch": 0.6,
-      "learning_rate": 0.00014735812133072405,
-      "loss": 0.5837,
+      "learning_rate": 4.911937377690802e-05,
+      "loss": 0.5661,
       "step": 750
     },
     {
       "epoch": 0.6,
-      "eval_accuracy": 0.6925373134328359,
-      "eval_loss": 0.6470033526420593,
-      "eval_runtime": 61.3513,
-      "eval_samples_per_second": 5.46,
-      "eval_steps_per_second": 1.369,
+      "eval_accuracy": 0.6835820895522388,
+      "eval_loss": 0.7564846277236938,
+      "eval_runtime": 56.4648,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 750
     },
     {
       "epoch": 0.64,
-      "learning_rate": 0.00013268101761252445,
-      "loss": 0.623,
+      "learning_rate": 4.422700587084149e-05,
+      "loss": 0.5675,
       "step": 800
     },
     {
       "epoch": 0.64,
-      "eval_accuracy": 0.6776119402985075,
-      "eval_loss": 0.7178429365158081,
-      "eval_runtime": 56.5122,
-      "eval_samples_per_second": 5.928,
-      "eval_steps_per_second": 1.486,
+      "eval_accuracy": 0.6417910447761194,
+      "eval_loss": 0.6759604215621948,
+      "eval_runtime": 56.4652,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 800
     },
     {
       "epoch": 0.68,
-      "learning_rate": 0.00011800391389432484,
-      "loss": 0.6285,
+      "learning_rate": 3.933463796477495e-05,
+      "loss": 0.5611,
       "step": 850
     },
     {
       "epoch": 0.68,
-      "eval_accuracy": 0.6805970149253732,
-      "eval_loss": 0.7012776732444763,
-      "eval_runtime": 61.0309,
-      "eval_samples_per_second": 5.489,
-      "eval_steps_per_second": 1.376,
+      "eval_accuracy": 0.6477611940298508,
+      "eval_loss": 0.6798116564750671,
+      "eval_runtime": 56.4714,
+      "eval_samples_per_second": 5.932,
+      "eval_steps_per_second": 1.487,
       "step": 850
     },
     {
       "epoch": 0.72,
-      "learning_rate": 0.00010332681017612522,
-      "loss": 0.609,
+      "learning_rate": 3.4442270058708414e-05,
+      "loss": 0.5664,
       "step": 900
     },
     {
       "epoch": 0.72,
       "eval_accuracy": 0.6656716417910448,
-      "eval_loss": 0.6416519284248352,
-      "eval_runtime": 56.5079,
-      "eval_samples_per_second": 5.928,
+      "eval_loss": 0.7185413837432861,
+      "eval_runtime": 56.4811,
+      "eval_samples_per_second": 5.931,
       "eval_steps_per_second": 1.487,
       "step": 900
     },
     {
       "epoch": 0.76,
-      "learning_rate": 8.864970645792562e-05,
-      "loss": 0.6,
+      "learning_rate": 2.9549902152641878e-05,
+      "loss": 0.5765,
       "step": 950
     },
     {
       "epoch": 0.76,
-      "eval_accuracy": 0.6835820895522388,
-      "eval_loss": 0.6628562808036804,
-      "eval_runtime": 61.1667,
-      "eval_samples_per_second": 5.477,
-      "eval_steps_per_second": 1.373,
+      "eval_accuracy": 0.6417910447761194,
+      "eval_loss": 0.6795146465301514,
+      "eval_runtime": 56.4589,
+      "eval_samples_per_second": 5.934,
+      "eval_steps_per_second": 1.488,
       "step": 950
     },
     {
       "epoch": 0.8,
-      "learning_rate": 7.397260273972601e-05,
-      "loss": 0.6142,
+      "learning_rate": 2.4657534246575342e-05,
+      "loss": 0.555,
       "step": 1000
     },
     {
       "epoch": 0.8,
-      "eval_accuracy": 0.6805970149253732,
-      "eval_loss": 0.635610044002533,
-      "eval_runtime": 61.6579,
-      "eval_samples_per_second": 5.433,
-      "eval_steps_per_second": 1.362,
+      "eval_accuracy": 0.6298507462686567,
+      "eval_loss": 0.677854061126709,
+      "eval_runtime": 56.4673,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 1000
     },
     {
       "epoch": 0.84,
-      "learning_rate": 5.929549902152641e-05,
-      "loss": 0.5855,
+      "learning_rate": 1.9765166340508805e-05,
+      "loss": 0.5578,
       "step": 1050
     },
     {
       "epoch": 0.84,
-      "eval_accuracy": 0.6955223880597015,
-      "eval_loss": 0.6755445003509521,
-      "eval_runtime": 61.4996,
-      "eval_samples_per_second": 5.447,
-      "eval_steps_per_second": 1.366,
+      "eval_accuracy": 0.6507462686567164,
+      "eval_loss": 0.6885824203491211,
+      "eval_runtime": 56.4615,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 1050
     },
     {
       "epoch": 0.88,
-      "learning_rate": 4.46183953033268e-05,
-      "loss": 0.5741,
+      "learning_rate": 1.4872798434442269e-05,
+      "loss": 0.5709,
       "step": 1100
     },
     {
       "epoch": 0.88,
-      "eval_accuracy": 0.6835820895522388,
-      "eval_loss": 0.6594328284263611,
-      "eval_runtime": 56.5065,
-      "eval_samples_per_second": 5.929,
-      "eval_steps_per_second": 1.487,
+      "eval_accuracy": 0.6447761194029851,
+      "eval_loss": 0.6935857534408569,
+      "eval_runtime": 56.4542,
+      "eval_samples_per_second": 5.934,
+      "eval_steps_per_second": 1.488,
       "step": 1100
     },
     {
       "epoch": 0.92,
-      "learning_rate": 2.99412915851272e-05,
-      "loss": 0.5918,
+      "learning_rate": 9.980430528375734e-06,
+      "loss": 0.5711,
       "step": 1150
     },
     {
       "epoch": 0.92,
-      "eval_accuracy": 0.6776119402985075,
-      "eval_loss": 0.6604845523834229,
-      "eval_runtime": 61.4144,
-      "eval_samples_per_second": 5.455,
-      "eval_steps_per_second": 1.368,
+      "eval_accuracy": 0.6388059701492538,
+      "eval_loss": 0.6908484101295471,
+      "eval_runtime": 56.4666,
+      "eval_samples_per_second": 5.933,
+      "eval_steps_per_second": 1.488,
       "step": 1150
     },
     {
       "epoch": 0.96,
-      "learning_rate": 1.526418786692759e-05,
-      "loss": 0.597,
+      "learning_rate": 5.088062622309197e-06,
+      "loss": 0.5597,
       "step": 1200
     },
     {
       "epoch": 0.96,
-      "eval_accuracy": 0.6925373134328359,
-      "eval_loss": 0.6844801902770996,
-      "eval_runtime": 61.5848,
-      "eval_samples_per_second": 5.44,
-      "eval_steps_per_second": 1.364,
+      "eval_accuracy": 0.6358208955223881,
+      "eval_loss": 0.6834951639175415,
+      "eval_runtime": 56.4792,
+      "eval_samples_per_second": 5.931,
+      "eval_steps_per_second": 1.487,
       "step": 1200
     }
   ],
   "logging_steps": 50,
   "max_steps": 1247,
   "num_train_epochs": 1,
-  "save_steps": 100,
-  "total_flos": 4.4953417384697856e+17,
+  "save_steps": 50,
+  "total_flos": 4.481413478232883e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1200/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e32e120c6d65576e6fe76fdf63d907877e27469586cc746d8cd6a71d07635ea9
+oid sha256:87bd30bd4b21b4a0c66f1fd545858ef6e2d33b5984a8d292e79011acc3affd6b
 size 4472