End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 89.98,
+    "eval_loss": 0.7423784136772156,
+    "eval_runtime": 31.8664,
+    "eval_samples": 892,
+    "eval_samples_per_second": 27.992,
+    "eval_steps_per_second": 3.515,
+    "eval_wer": 0.6190132717830352,
+    "train_loss": 1.793763725845902,
+    "train_runtime": 8224.5964,
+    "train_samples": 1941,
+    "train_samples_per_second": 21.24,
+    "train_steps_per_second": 0.328
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 89.98,
+    "eval_loss": 0.7423784136772156,
+    "eval_runtime": 31.8664,
+    "eval_samples": 892,
+    "eval_samples_per_second": 27.992,
+    "eval_steps_per_second": 3.515,
+    "eval_wer": 0.6190132717830352
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 89.98,
+    "train_loss": 1.793763725845902,
+    "train_runtime": 8224.5964,
+    "train_samples": 1941,
+    "train_samples_per_second": 21.24,
+    "train_steps_per_second": 0.328
+}

trainer_state.json

@@ -0,0 +1,232 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 89.98360655737704,
+  "global_step": 2700,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 3.33,
+      "learning_rate": 1.485e-05,
+      "loss": 11.2315,
+      "step": 100
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 2.985e-05,
+      "loss": 4.0404,
+      "step": 200
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 4.484999999999999e-05,
+      "loss": 3.056,
+      "step": 300
+    },
+    {
+      "epoch": 13.33,
+      "learning_rate": 5.985e-05,
+      "loss": 2.8167,
+      "step": 400
+    },
+    {
+      "epoch": 16.66,
+      "learning_rate": 7.484999999999999e-05,
+      "loss": 2.721,
+      "step": 500
+    },
+    {
+      "epoch": 16.66,
+      "eval_loss": 2.6876425743103027,
+      "eval_runtime": 32.088,
+      "eval_samples_per_second": 27.799,
+      "eval_steps_per_second": 3.49,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 19.98,
+      "learning_rate": 8.984999999999999e-05,
+      "loss": 2.6493,
+      "step": 600
+    },
+    {
+      "epoch": 23.33,
+      "learning_rate": 0.00010484999999999999,
+      "loss": 1.9825,
+      "step": 700
+    },
+    {
+      "epoch": 26.66,
+      "learning_rate": 0.00011985,
+      "loss": 1.4345,
+      "step": 800
+    },
+    {
+      "epoch": 29.98,
+      "learning_rate": 0.00013485,
+      "loss": 1.3429,
+      "step": 900
+    },
+    {
+      "epoch": 33.33,
+      "learning_rate": 0.00014954999999999998,
+      "loss": 1.2944,
+      "step": 1000
+    },
+    {
+      "epoch": 33.33,
+      "eval_loss": 0.6482492089271545,
+      "eval_runtime": 32.0457,
+      "eval_samples_per_second": 27.835,
+      "eval_steps_per_second": 3.495,
+      "eval_wer": 0.7135025966532026,
+      "step": 1000
+    },
+    {
+      "epoch": 36.66,
+      "learning_rate": 0.00016455,
+      "loss": 1.2347,
+      "step": 1100
+    },
+    {
+      "epoch": 39.98,
+      "learning_rate": 0.00017955,
+      "loss": 1.1858,
+      "step": 1200
+    },
+    {
+      "epoch": 43.33,
+      "learning_rate": 0.00019454999999999999,
+      "loss": 1.1475,
+      "step": 1300
+    },
+    {
+      "epoch": 46.66,
+      "learning_rate": 0.00020955,
+      "loss": 1.0875,
+      "step": 1400
+    },
+    {
+      "epoch": 49.98,
+      "learning_rate": 0.00022455,
+      "loss": 1.0515,
+      "step": 1500
+    },
+    {
+      "epoch": 49.98,
+      "eval_loss": 0.6644838452339172,
+      "eval_runtime": 31.726,
+      "eval_samples_per_second": 28.116,
+      "eval_steps_per_second": 3.53,
+      "eval_wer": 0.6754183496826313,
+      "step": 1500
+    },
+    {
+      "epoch": 53.33,
+      "learning_rate": 0.00023954999999999997,
+      "loss": 1.0217,
+      "step": 1600
+    },
+    {
+      "epoch": 56.66,
+      "learning_rate": 0.00025455,
+      "loss": 0.9991,
+      "step": 1700
+    },
+    {
+      "epoch": 59.98,
+      "learning_rate": 0.00026954999999999997,
+      "loss": 0.9918,
+      "step": 1800
+    },
+    {
+      "epoch": 63.33,
+      "learning_rate": 0.00028455,
+      "loss": 0.9552,
+      "step": 1900
+    },
+    {
+      "epoch": 66.66,
+      "learning_rate": 0.00029955,
+      "loss": 0.9153,
+      "step": 2000
+    },
+    {
+      "epoch": 66.66,
+      "eval_loss": 0.7648739814758301,
+      "eval_runtime": 31.6976,
+      "eval_samples_per_second": 28.141,
+      "eval_steps_per_second": 3.533,
+      "eval_wer": 0.693594922100404,
+      "step": 2000
+    },
+    {
+      "epoch": 69.98,
+      "learning_rate": 0.0002584285714285714,
+      "loss": 0.8827,
+      "step": 2100
+    },
+    {
+      "epoch": 73.33,
+      "learning_rate": 0.00021557142857142855,
+      "loss": 0.8474,
+      "step": 2200
+    },
+    {
+      "epoch": 76.66,
+      "learning_rate": 0.0001727142857142857,
+      "loss": 0.8016,
+      "step": 2300
+    },
+    {
+      "epoch": 79.98,
+      "learning_rate": 0.00012985714285714285,
+      "loss": 0.7467,
+      "step": 2400
+    },
+    {
+      "epoch": 83.33,
+      "learning_rate": 8.699999999999999e-05,
+      "loss": 0.7056,
+      "step": 2500
+    },
+    {
+      "epoch": 83.33,
+      "eval_loss": 0.8047966361045837,
+      "eval_runtime": 31.4528,
+      "eval_samples_per_second": 28.36,
+      "eval_steps_per_second": 3.561,
+      "eval_wer": 0.6755626081938835,
+      "step": 2500
+    },
+    {
+      "epoch": 86.66,
+      "learning_rate": 4.414285714285714e-05,
+      "loss": 0.6637,
+      "step": 2600
+    },
+    {
+      "epoch": 89.98,
+      "learning_rate": 1.2857142857142856e-06,
+      "loss": 0.6249,
+      "step": 2700
+    },
+    {
+      "epoch": 89.98,
+      "step": 2700,
+      "total_flos": 2.311805909026184e+19,
+      "train_loss": 1.793763725845902,
+      "train_runtime": 8224.5964,
+      "train_samples_per_second": 21.24,
+      "train_steps_per_second": 0.328
+    }
+  ],
+  "max_steps": 2700,
+  "num_train_epochs": 90,
+  "total_flos": 2.311805909026184e+19,
+  "trial_name": null,
+  "trial_params": null
+}