End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.31550145149230957,
+    "eval_runtime": 26.3335,
+    "eval_samples": 421,
+    "eval_samples_per_second": 15.987,
+    "eval_steps_per_second": 15.987,
+    "eval_wer": 0.316162109375,
+    "train_loss": 1.9088435792922973,
+    "train_runtime": 7870.625,
+    "train_samples": 1010,
+    "train_samples_per_second": 12.833,
+    "train_steps_per_second": 0.407
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 100.0,
+    "eval_loss": 0.31550145149230957,
+    "eval_runtime": 26.3335,
+    "eval_samples": 421,
+    "eval_samples_per_second": 15.987,
+    "eval_steps_per_second": 15.987,
+    "eval_wer": 0.316162109375
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a2005bc1e2b51942c72e991d7a2c3a2e4cb82773c9a598e6309ea89cca2b294
+oid sha256:56ab36bf51bc16b26d7447d684a1f1b88e400a1aca4d198511e063e8fb481b8b
 size 1262112241

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 100.0,
+    "train_loss": 1.9088435792922973,
+    "train_runtime": 7870.625,
+    "train_samples": 1010,
+    "train_samples_per_second": 12.833,
+    "train_steps_per_second": 0.407
+}

trainer_state.json

@@ -0,0 +1,271 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 100.0,
+  "global_step": 3200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.3579999999999999e-05,
+      "loss": 14.122,
+      "step": 100
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 2.758e-05,
+      "loss": 5.0536,
+      "step": 200
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 4.157999999999999e-05,
+      "loss": 3.5787,
+      "step": 300
+    },
+    {
+      "epoch": 12.5,
+      "learning_rate": 5.558e-05,
+      "loss": 3.0454,
+      "step": 400
+    },
+    {
+      "epoch": 15.62,
+      "learning_rate": 6.958e-05,
+      "loss": 2.9556,
+      "step": 500
+    },
+    {
+      "epoch": 15.62,
+      "eval_loss": 2.929969549179077,
+      "eval_runtime": 26.3053,
+      "eval_samples_per_second": 16.004,
+      "eval_steps_per_second": 16.004,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 18.75,
+      "learning_rate": 6.748518518518517e-05,
+      "loss": 2.8989,
+      "step": 600
+    },
+    {
+      "epoch": 21.88,
+      "learning_rate": 6.489259259259258e-05,
+      "loss": 2.8712,
+      "step": 700
+    },
+    {
+      "epoch": 25.0,
+      "learning_rate": 6.23e-05,
+      "loss": 2.8548,
+      "step": 800
+    },
+    {
+      "epoch": 28.12,
+      "learning_rate": 5.97074074074074e-05,
+      "loss": 2.6412,
+      "step": 900
+    },
+    {
+      "epoch": 31.25,
+      "learning_rate": 5.711481481481481e-05,
+      "loss": 1.7874,
+      "step": 1000
+    },
+    {
+      "epoch": 31.25,
+      "eval_loss": 0.7566312551498413,
+      "eval_runtime": 26.7283,
+      "eval_samples_per_second": 15.751,
+      "eval_steps_per_second": 15.751,
+      "eval_wer": 0.65087890625,
+      "step": 1000
+    },
+    {
+      "epoch": 34.38,
+      "learning_rate": 5.4522222222222216e-05,
+      "loss": 1.3943,
+      "step": 1100
+    },
+    {
+      "epoch": 37.5,
+      "learning_rate": 5.1929629629629626e-05,
+      "loss": 1.2438,
+      "step": 1200
+    },
+    {
+      "epoch": 40.62,
+      "learning_rate": 4.933703703703703e-05,
+      "loss": 1.1248,
+      "step": 1300
+    },
+    {
+      "epoch": 43.75,
+      "learning_rate": 4.674444444444444e-05,
+      "loss": 1.0708,
+      "step": 1400
+    },
+    {
+      "epoch": 46.88,
+      "learning_rate": 4.4151851851851846e-05,
+      "loss": 1.0131,
+      "step": 1500
+    },
+    {
+      "epoch": 46.88,
+      "eval_loss": 0.36714494228363037,
+      "eval_runtime": 25.9364,
+      "eval_samples_per_second": 16.232,
+      "eval_steps_per_second": 16.232,
+      "eval_wer": 0.3828125,
+      "step": 1500
+    },
+    {
+      "epoch": 50.0,
+      "learning_rate": 4.1559259259259257e-05,
+      "loss": 0.9624,
+      "step": 1600
+    },
+    {
+      "epoch": 53.12,
+      "learning_rate": 3.896666666666666e-05,
+      "loss": 0.9292,
+      "step": 1700
+    },
+    {
+      "epoch": 56.25,
+      "learning_rate": 3.637407407407407e-05,
+      "loss": 0.899,
+      "step": 1800
+    },
+    {
+      "epoch": 59.38,
+      "learning_rate": 3.3781481481481476e-05,
+      "loss": 0.8697,
+      "step": 1900
+    },
+    {
+      "epoch": 62.5,
+      "learning_rate": 3.1188888888888887e-05,
+      "loss": 0.8439,
+      "step": 2000
+    },
+    {
+      "epoch": 62.5,
+      "eval_loss": 0.33496010303497314,
+      "eval_runtime": 26.254,
+      "eval_samples_per_second": 16.036,
+      "eval_steps_per_second": 16.036,
+      "eval_wer": 0.341552734375,
+      "step": 2000
+    },
+    {
+      "epoch": 65.62,
+      "learning_rate": 2.8596296296296294e-05,
+      "loss": 0.8319,
+      "step": 2100
+    },
+    {
+      "epoch": 68.75,
+      "learning_rate": 2.60037037037037e-05,
+      "loss": 0.808,
+      "step": 2200
+    },
+    {
+      "epoch": 71.88,
+      "learning_rate": 2.341111111111111e-05,
+      "loss": 0.7916,
+      "step": 2300
+    },
+    {
+      "epoch": 75.0,
+      "learning_rate": 2.0818518518518517e-05,
+      "loss": 0.7731,
+      "step": 2400
+    },
+    {
+      "epoch": 78.12,
+      "learning_rate": 1.8225925925925924e-05,
+      "loss": 0.7502,
+      "step": 2500
+    },
+    {
+      "epoch": 78.12,
+      "eval_loss": 0.3154859244823456,
+      "eval_runtime": 26.713,
+      "eval_samples_per_second": 15.76,
+      "eval_steps_per_second": 15.76,
+      "eval_wer": 0.32958984375,
+      "step": 2500
+    },
+    {
+      "epoch": 81.25,
+      "learning_rate": 1.563333333333333e-05,
+      "loss": 0.7307,
+      "step": 2600
+    },
+    {
+      "epoch": 84.38,
+      "learning_rate": 1.3040740740740739e-05,
+      "loss": 0.7262,
+      "step": 2700
+    },
+    {
+      "epoch": 87.5,
+      "learning_rate": 1.0448148148148147e-05,
+      "loss": 0.7154,
+      "step": 2800
+    },
+    {
+      "epoch": 90.62,
+      "learning_rate": 7.855555555555554e-06,
+      "loss": 0.7092,
+      "step": 2900
+    },
+    {
+      "epoch": 93.75,
+      "learning_rate": 5.2629629629629625e-06,
+      "loss": 0.7093,
+      "step": 3000
+    },
+    {
+      "epoch": 93.75,
+      "eval_loss": 0.31820571422576904,
+      "eval_runtime": 27.0308,
+      "eval_samples_per_second": 15.575,
+      "eval_steps_per_second": 15.575,
+      "eval_wer": 0.318603515625,
+      "step": 3000
+    },
+    {
+      "epoch": 96.88,
+      "learning_rate": 2.67037037037037e-06,
+      "loss": 0.6896,
+      "step": 3100
+    },
+    {
+      "epoch": 100.0,
+      "learning_rate": 7.777777777777777e-08,
+      "loss": 0.688,
+      "step": 3200
+    },
+    {
+      "epoch": 100.0,
+      "step": 3200,
+      "total_flos": 2.067522226545355e+19,
+      "train_loss": 1.9088435792922973,
+      "train_runtime": 7870.625,
+      "train_samples_per_second": 12.833,
+      "train_steps_per_second": 0.407
+    }
+  ],
+  "max_steps": 3200,
+  "num_train_epochs": 100,
+  "total_flos": 2.067522226545355e+19,
+  "trial_name": null,
+  "trial_params": null
+}