End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 50.0,
+    "eval_loss": 0.19587931036949158,
+    "eval_runtime": 168.5422,
+    "eval_samples": 4686,
+    "eval_samples_per_second": 27.803,
+    "eval_steps_per_second": 1.738,
+    "eval_wer": 0.24543618426163288,
+    "train_loss": 1.2943066022937424,
+    "train_runtime": 28854.2781,
+    "train_samples": 13247,
+    "train_samples_per_second": 22.955,
+    "train_steps_per_second": 0.717
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 50.0,
+    "eval_loss": 0.19587931036949158,
+    "eval_runtime": 168.5422,
+    "eval_samples": 4686,
+    "eval_samples_per_second": 27.803,
+    "eval_steps_per_second": 1.738,
+    "eval_wer": 0.24543618426163288
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4a20a063e02613c3a6229c458c2b336acaa3c046e998b5a3f3ecf0bd5146c5e
+oid sha256:25f144236d73b276af60997ba3e53e23dab0fca4dc081a995b0a6dfef078e4dd
 size 1262145073

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 50.0,
+    "train_loss": 1.2943066022937424,
+    "train_runtime": 28854.2781,
+    "train_samples": 13247,
+    "train_samples_per_second": 22.955,
+    "train_steps_per_second": 0.717
+}

trainer_state.json

@@ -0,0 +1,190 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 50.0,
+  "global_step": 20700,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.74475e-05,
+      "loss": 7.0856,
+      "step": 1000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 3.49475e-05,
+      "loss": 2.771,
+      "step": 2000
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 5.243e-05,
+      "loss": 1.4051,
+      "step": 3000
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 6.99125e-05,
+      "loss": 1.173,
+      "step": 4000
+    },
+    {
+      "epoch": 9.66,
+      "eval_loss": 0.29204532504081726,
+      "eval_runtime": 168.5565,
+      "eval_samples_per_second": 27.801,
+      "eval_steps_per_second": 1.738,
+      "eval_wer": 0.36075506874854346,
+      "step": 4000
+    },
+    {
+      "epoch": 12.08,
+      "learning_rate": 6.582934131736526e-05,
+      "loss": 1.082,
+      "step": 5000
+    },
+    {
+      "epoch": 14.49,
+      "learning_rate": 6.164191616766467e-05,
+      "loss": 1.0256,
+      "step": 6000
+    },
+    {
+      "epoch": 16.91,
+      "learning_rate": 5.74502994011976e-05,
+      "loss": 0.9779,
+      "step": 7000
+    },
+    {
+      "epoch": 19.32,
+      "learning_rate": 5.3262874251497e-05,
+      "loss": 0.9433,
+      "step": 8000
+    },
+    {
+      "epoch": 19.32,
+      "eval_loss": 0.23364943265914917,
+      "eval_runtime": 167.8504,
+      "eval_samples_per_second": 27.918,
+      "eval_steps_per_second": 1.746,
+      "eval_wer": 0.30261011419249595,
+      "step": 8000
+    },
+    {
+      "epoch": 21.74,
+      "learning_rate": 4.907125748502993e-05,
+      "loss": 0.9211,
+      "step": 9000
+    },
+    {
+      "epoch": 24.15,
+      "learning_rate": 4.48880239520958e-05,
+      "loss": 0.8969,
+      "step": 10000
+    },
+    {
+      "epoch": 26.57,
+      "learning_rate": 4.069640718562874e-05,
+      "loss": 0.8727,
+      "step": 11000
+    },
+    {
+      "epoch": 28.99,
+      "learning_rate": 3.6504790419161677e-05,
+      "loss": 0.8552,
+      "step": 12000
+    },
+    {
+      "epoch": 28.99,
+      "eval_loss": 0.22206313908100128,
+      "eval_runtime": 169.6004,
+      "eval_samples_per_second": 27.63,
+      "eval_steps_per_second": 1.728,
+      "eval_wer": 0.27988813796317874,
+      "step": 12000
+    },
+    {
+      "epoch": 31.4,
+      "learning_rate": 3.231317365269461e-05,
+      "loss": 0.8401,
+      "step": 13000
+    },
+    {
+      "epoch": 33.82,
+      "learning_rate": 2.8121556886227542e-05,
+      "loss": 0.8178,
+      "step": 14000
+    },
+    {
+      "epoch": 36.23,
+      "learning_rate": 2.3934131736526942e-05,
+      "loss": 0.8002,
+      "step": 15000
+    },
+    {
+      "epoch": 38.65,
+      "learning_rate": 1.974251497005988e-05,
+      "loss": 0.7863,
+      "step": 16000
+    },
+    {
+      "epoch": 38.65,
+      "eval_loss": 0.1952943205833435,
+      "eval_runtime": 169.2585,
+      "eval_samples_per_second": 27.685,
+      "eval_steps_per_second": 1.731,
+      "eval_wer": 0.24792200730210517,
+      "step": 16000
+    },
+    {
+      "epoch": 41.06,
+      "learning_rate": 1.555508982035928e-05,
+      "loss": 0.778,
+      "step": 17000
+    },
+    {
+      "epoch": 43.48,
+      "learning_rate": 1.1367664670658683e-05,
+      "loss": 0.7607,
+      "step": 18000
+    },
+    {
+      "epoch": 45.89,
+      "learning_rate": 7.180239520958083e-06,
+      "loss": 0.7481,
+      "step": 19000
+    },
+    {
+      "epoch": 48.31,
+      "learning_rate": 2.988622754491018e-06,
+      "loss": 0.7365,
+      "step": 20000
+    },
+    {
+      "epoch": 48.31,
+      "eval_loss": 0.1967712938785553,
+      "eval_runtime": 167.9953,
+      "eval_samples_per_second": 27.894,
+      "eval_steps_per_second": 1.744,
+      "eval_wer": 0.24489241047152957,
+      "step": 20000
+    },
+    {
+      "epoch": 50.0,
+      "step": 20700,
+      "total_flos": 7.668153416081444e+19,
+      "train_loss": 1.2943066022937424,
+      "train_runtime": 28854.2781,
+      "train_samples_per_second": 22.955,
+      "train_steps_per_second": 0.717
+    }
+  ],
+  "max_steps": 20700,
+  "num_train_epochs": 50,
+  "total_flos": 7.668153416081444e+19,
+  "trial_name": null,
+  "trial_params": null
+}