End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 4.0,
+    "eval_accuracy": 0.9036111798169676,
+    "eval_loss": 0.43492814898490906,
+    "eval_runtime": 86.0866,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 469.643,
+    "eval_steps_per_second": 14.683,
+    "train_loss": 0.16138914258944828,
+    "train_runtime": 10416.0445,
+    "train_samples": 363846,
+    "train_samples_per_second": 139.725,
+    "train_steps_per_second": 1.091
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 4.0,
+    "eval_accuracy": 0.9036111798169676,
+    "eval_loss": 0.43492814898490906,
+    "eval_runtime": 86.0866,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 469.643,
+    "eval_steps_per_second": 14.683
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 4.0,
+    "train_loss": 0.16138914258944828,
+    "train_runtime": 10416.0445,
+    "train_samples": 363846,
+    "train_samples_per_second": 139.725,
+    "train_steps_per_second": 1.091
+}

trainer_state.json

@@ -0,0 +1,193 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.999120531199156,
+  "global_step": 11368,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.780524278676988e-05,
+      "loss": 0.3883,
+      "step": 500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.56060872624912e-05,
+      "loss": 0.3251,
+      "step": 1000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.340693173821253e-05,
+      "loss": 0.2984,
+      "step": 1500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.121217452498241e-05,
+      "loss": 0.2895,
+      "step": 2000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.9013019000703735e-05,
+      "loss": 0.2782,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8864209745238684,
+      "eval_loss": 0.26690077781677246,
+      "eval_runtime": 86.4948,
+      "eval_samples_per_second": 467.427,
+      "eval_steps_per_second": 14.614,
+      "step": 2842
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.6813863476425054e-05,
+      "loss": 0.2481,
+      "step": 3000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.461470795214638e-05,
+      "loss": 0.1938,
+      "step": 3500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.24155524278677e-05,
+      "loss": 0.1904,
+      "step": 4000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.0220795214637583e-05,
+      "loss": 0.1858,
+      "step": 4500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.8021639690358902e-05,
+      "loss": 0.1871,
+      "step": 5000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.5822484166080225e-05,
+      "loss": 0.1866,
+      "step": 5500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.899060103883255,
+      "eval_loss": 0.24605706334114075,
+      "eval_runtime": 86.4559,
+      "eval_samples_per_second": 467.637,
+      "eval_steps_per_second": 14.62,
+      "step": 5685
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.3623328641801548e-05,
+      "loss": 0.1354,
+      "step": 6000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.1428571428571428e-05,
+      "loss": 0.1054,
+      "step": 6500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.923381421534131e-05,
+      "loss": 0.1011,
+      "step": 7000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.703465869106263e-05,
+      "loss": 0.1019,
+      "step": 7500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.4835503166783957e-05,
+      "loss": 0.1007,
+      "step": 8000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.2636347642505278e-05,
+      "loss": 0.0989,
+      "step": 8500
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9031412317585951,
+      "eval_loss": 0.3130354583263397,
+      "eval_runtime": 86.6189,
+      "eval_samples_per_second": 466.757,
+      "eval_steps_per_second": 14.593,
+      "step": 8527
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.0437192118226603e-05,
+      "loss": 0.0466,
+      "step": 9000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 8.238036593947925e-06,
+      "loss": 0.0474,
+      "step": 9500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.038881069669247e-06,
+      "loss": 0.0431,
+      "step": 10000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.839725545390571e-06,
+      "loss": 0.0445,
+      "step": 10500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.640570021111893e-06,
+      "loss": 0.0416,
+      "step": 11000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9036111798169676,
+      "eval_loss": 0.43492814898490906,
+      "eval_runtime": 86.4404,
+      "eval_samples_per_second": 467.721,
+      "eval_steps_per_second": 14.623,
+      "step": 11368
+    },
+    {
+      "epoch": 4.0,
+      "step": 11368,
+      "total_flos": 3.802061080709038e+17,
+      "train_loss": 0.16138914258944828,
+      "train_runtime": 10416.0445,
+      "train_samples_per_second": 139.725,
+      "train_steps_per_second": 1.091
+    }
+  ],
+  "max_steps": 11368,
+  "num_train_epochs": 4,
+  "total_flos": 3.802061080709038e+17,
+  "trial_name": null,
+  "trial_params": null
+}