End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.817560613155365,
+    "eval_runtime": 10.3871,
+    "eval_samples": 213,
+    "eval_samples_per_second": 20.506,
+    "eval_steps_per_second": 1.348,
+    "eval_wer": 0.5818422796554009,
+    "train_loss": 1.675173250834147,
+    "train_runtime": 5040.7553,
+    "train_samples": 753,
+    "train_samples_per_second": 14.938,
+    "train_steps_per_second": 0.476
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.817560613155365,
+    "eval_runtime": 10.3871,
+    "eval_samples": 213,
+    "eval_samples_per_second": 20.506,
+    "eval_steps_per_second": 1.348,
+    "eval_wer": 0.5818422796554009
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 100.0,
+    "train_loss": 1.675173250834147,
+    "train_runtime": 5040.7553,
+    "train_samples": 753,
+    "train_samples_per_second": 14.938,
+    "train_steps_per_second": 0.476
+}

trainer_state.json

@@ -0,0 +1,205 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 100.0,
+  "global_step": 2400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 4.17,
+      "learning_rate": 5.82e-05,
+      "loss": 11.8246,
+      "step": 100
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 0.0001182,
+      "loss": 3.9245,
+      "step": 200
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 0.00017819999999999997,
+      "loss": 3.3755,
+      "step": 300
+    },
+    {
+      "epoch": 16.67,
+      "learning_rate": 0.0002382,
+      "loss": 3.2775,
+      "step": 400
+    },
+    {
+      "epoch": 20.83,
+      "learning_rate": 0.0002982,
+      "loss": 2.3957,
+      "step": 500
+    },
+    {
+      "epoch": 20.83,
+      "eval_loss": 1.0925403833389282,
+      "eval_runtime": 10.3822,
+      "eval_samples_per_second": 20.516,
+      "eval_steps_per_second": 1.348,
+      "eval_wer": 0.8111332007952287,
+      "step": 500
+    },
+    {
+      "epoch": 25.0,
+      "learning_rate": 0.0002846842105263158,
+      "loss": 1.5517,
+      "step": 600
+    },
+    {
+      "epoch": 29.17,
+      "learning_rate": 0.0002688947368421052,
+      "loss": 1.3096,
+      "step": 700
+    },
+    {
+      "epoch": 33.33,
+      "learning_rate": 0.0002531052631578947,
+      "loss": 1.2066,
+      "step": 800
+    },
+    {
+      "epoch": 37.5,
+      "learning_rate": 0.0002373157894736842,
+      "loss": 1.1194,
+      "step": 900
+    },
+    {
+      "epoch": 41.67,
+      "learning_rate": 0.00022152631578947366,
+      "loss": 1.0351,
+      "step": 1000
+    },
+    {
+      "epoch": 41.67,
+      "eval_loss": 0.7836905717849731,
+      "eval_runtime": 11.7781,
+      "eval_samples_per_second": 18.084,
+      "eval_steps_per_second": 1.189,
+      "eval_wer": 0.6573889993373094,
+      "step": 1000
+    },
+    {
+      "epoch": 45.83,
+      "learning_rate": 0.00020573684210526314,
+      "loss": 0.955,
+      "step": 1100
+    },
+    {
+      "epoch": 50.0,
+      "learning_rate": 0.00018994736842105263,
+      "loss": 0.8928,
+      "step": 1200
+    },
+    {
+      "epoch": 54.17,
+      "learning_rate": 0.00017415789473684208,
+      "loss": 0.8239,
+      "step": 1300
+    },
+    {
+      "epoch": 58.33,
+      "learning_rate": 0.00015836842105263158,
+      "loss": 0.7782,
+      "step": 1400
+    },
+    {
+      "epoch": 62.5,
+      "learning_rate": 0.00014257894736842105,
+      "loss": 0.7396,
+      "step": 1500
+    },
+    {
+      "epoch": 62.5,
+      "eval_loss": 0.7674338221549988,
+      "eval_runtime": 10.3819,
+      "eval_samples_per_second": 20.516,
+      "eval_steps_per_second": 1.348,
+      "eval_wer": 0.6083499005964215,
+      "step": 1500
+    },
+    {
+      "epoch": 66.67,
+      "learning_rate": 0.00012678947368421052,
+      "loss": 0.6811,
+      "step": 1600
+    },
+    {
+      "epoch": 70.83,
+      "learning_rate": 0.00011099999999999999,
+      "loss": 0.6438,
+      "step": 1700
+    },
+    {
+      "epoch": 75.0,
+      "learning_rate": 9.521052631578946e-05,
+      "loss": 0.6074,
+      "step": 1800
+    },
+    {
+      "epoch": 79.17,
+      "learning_rate": 7.942105263157894e-05,
+      "loss": 0.5791,
+      "step": 1900
+    },
+    {
+      "epoch": 83.33,
+      "learning_rate": 6.363157894736841e-05,
+      "loss": 0.5385,
+      "step": 2000
+    },
+    {
+      "epoch": 83.33,
+      "eval_loss": 0.801542341709137,
+      "eval_runtime": 10.3837,
+      "eval_samples_per_second": 20.513,
+      "eval_steps_per_second": 1.348,
+      "eval_wer": 0.5811795891318754,
+      "step": 2000
+    },
+    {
+      "epoch": 87.5,
+      "learning_rate": 4.784210526315789e-05,
+      "loss": 0.5167,
+      "step": 2100
+    },
+    {
+      "epoch": 91.67,
+      "learning_rate": 3.205263157894736e-05,
+      "loss": 0.495,
+      "step": 2200
+    },
+    {
+      "epoch": 95.83,
+      "learning_rate": 1.626315789473684e-05,
+      "loss": 0.4769,
+      "step": 2300
+    },
+    {
+      "epoch": 100.0,
+      "learning_rate": 4.7368421052631574e-07,
+      "loss": 0.4558,
+      "step": 2400
+    },
+    {
+      "epoch": 100.0,
+      "step": 2400,
+      "total_flos": 1.3881399754298573e+19,
+      "train_loss": 1.675173250834147,
+      "train_runtime": 5040.7553,
+      "train_samples_per_second": 14.938,
+      "train_steps_per_second": 0.476
+    }
+  ],
+  "max_steps": 2400,
+  "num_train_epochs": 100,
+  "total_flos": 1.3881399754298573e+19,
+  "trial_name": null,
+  "trial_params": null
+}