End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.6483336687088013,
+    "eval_runtime": 14.4389,
+    "eval_samples": 306,
+    "eval_samples_per_second": 21.193,
+    "eval_steps_per_second": 1.385,
+    "eval_wer": 0.6048937023666265,
+    "train_loss": 1.7937395373257723,
+    "train_runtime": 4833.3951,
+    "train_samples": 698,
+    "train_samples_per_second": 14.441,
+    "train_steps_per_second": 0.455
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.6483336687088013,
+    "eval_runtime": 14.4389,
+    "eval_samples": 306,
+    "eval_samples_per_second": 21.193,
+    "eval_steps_per_second": 1.385,
+    "eval_wer": 0.6048937023666265
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 100.0,
+    "train_loss": 1.7937395373257723,
+    "train_runtime": 4833.3951,
+    "train_samples": 698,
+    "train_samples_per_second": 14.441,
+    "train_steps_per_second": 0.455
+}

trainer_state.json

@@ -0,0 +1,193 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 100.0,
+  "global_step": 2200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 4.55,
+      "learning_rate": 5.88e-05,
+      "loss": 9.8087,
+      "step": 100
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 0.0001188,
+      "loss": 3.8006,
+      "step": 200
+    },
+    {
+      "epoch": 13.64,
+      "learning_rate": 0.00017879999999999998,
+      "loss": 3.324,
+      "step": 300
+    },
+    {
+      "epoch": 18.18,
+      "learning_rate": 0.0002388,
+      "loss": 3.2068,
+      "step": 400
+    },
+    {
+      "epoch": 22.73,
+      "learning_rate": 0.0002988,
+      "loss": 2.671,
+      "step": 500
+    },
+    {
+      "epoch": 22.73,
+      "eval_loss": 1.3617517948150635,
+      "eval_runtime": 14.6151,
+      "eval_samples_per_second": 20.937,
+      "eval_steps_per_second": 1.368,
+      "eval_wer": 0.9498596068993181,
+      "step": 500
+    },
+    {
+      "epoch": 27.27,
+      "learning_rate": 0.00028270588235294116,
+      "loss": 1.8697,
+      "step": 600
+    },
+    {
+      "epoch": 31.82,
+      "learning_rate": 0.00026505882352941175,
+      "loss": 1.5472,
+      "step": 700
+    },
+    {
+      "epoch": 36.36,
+      "learning_rate": 0.00024741176470588234,
+      "loss": 1.3942,
+      "step": 800
+    },
+    {
+      "epoch": 40.91,
+      "learning_rate": 0.00022976470588235293,
+      "loss": 1.2455,
+      "step": 900
+    },
+    {
+      "epoch": 45.45,
+      "learning_rate": 0.00021211764705882352,
+      "loss": 1.1599,
+      "step": 1000
+    },
+    {
+      "epoch": 45.45,
+      "eval_loss": 0.6329962015151978,
+      "eval_runtime": 14.7493,
+      "eval_samples_per_second": 20.747,
+      "eval_steps_per_second": 1.356,
+      "eval_wer": 0.6626554352186121,
+      "step": 1000
+    },
+    {
+      "epoch": 50.0,
+      "learning_rate": 0.00019447058823529408,
+      "loss": 1.0693,
+      "step": 1100
+    },
+    {
+      "epoch": 54.55,
+      "learning_rate": 0.00017682352941176467,
+      "loss": 0.9951,
+      "step": 1200
+    },
+    {
+      "epoch": 59.09,
+      "learning_rate": 0.00015917647058823529,
+      "loss": 0.9336,
+      "step": 1300
+    },
+    {
+      "epoch": 63.64,
+      "learning_rate": 0.00014152941176470588,
+      "loss": 0.8805,
+      "step": 1400
+    },
+    {
+      "epoch": 68.18,
+      "learning_rate": 0.00012388235294117647,
+      "loss": 0.8252,
+      "step": 1500
+    },
+    {
+      "epoch": 68.18,
+      "eval_loss": 0.6226372122764587,
+      "eval_runtime": 15.3292,
+      "eval_samples_per_second": 19.962,
+      "eval_steps_per_second": 1.305,
+      "eval_wer": 0.6425992779783394,
+      "step": 1500
+    },
+    {
+      "epoch": 72.73,
+      "learning_rate": 0.00010623529411764705,
+      "loss": 0.7796,
+      "step": 1600
+    },
+    {
+      "epoch": 77.27,
+      "learning_rate": 8.858823529411763e-05,
+      "loss": 0.733,
+      "step": 1700
+    },
+    {
+      "epoch": 81.82,
+      "learning_rate": 7.094117647058823e-05,
+      "loss": 0.7071,
+      "step": 1800
+    },
+    {
+      "epoch": 86.36,
+      "learning_rate": 5.329411764705882e-05,
+      "loss": 0.6584,
+      "step": 1900
+    },
+    {
+      "epoch": 90.91,
+      "learning_rate": 3.564705882352941e-05,
+      "loss": 0.6424,
+      "step": 2000
+    },
+    {
+      "epoch": 90.91,
+      "eval_loss": 0.6359109878540039,
+      "eval_runtime": 14.8772,
+      "eval_samples_per_second": 20.568,
+      "eval_steps_per_second": 1.344,
+      "eval_wer": 0.6040914560770156,
+      "step": 2000
+    },
+    {
+      "epoch": 95.45,
+      "learning_rate": 1.7999999999999997e-05,
+      "loss": 0.6096,
+      "step": 2100
+    },
+    {
+      "epoch": 100.0,
+      "learning_rate": 3.5294117647058817e-07,
+      "loss": 0.601,
+      "step": 2200
+    },
+    {
+      "epoch": 100.0,
+      "step": 2200,
+      "total_flos": 1.307344883713109e+19,
+      "train_loss": 1.7937395373257723,
+      "train_runtime": 4833.3951,
+      "train_samples_per_second": 14.441,
+      "train_steps_per_second": 0.455
+    }
+  ],
+  "max_steps": 2200,
+  "num_train_epochs": 100,
+  "total_flos": 1.307344883713109e+19,
+  "trial_name": null,
+  "trial_params": null
+}