End of training

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 4.0,
+    "eval_accuracy": 0.9076675735839723,
+    "eval_loss": 0.3249448239803314,
+    "eval_runtime": 126.4373,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 319.763,
+    "eval_steps_per_second": 4.999,
+    "train_loss": 0.17184808974027801,
+    "train_runtime": 15393.9025,
+    "train_samples": 363846,
+    "train_samples_per_second": 94.543,
+    "train_steps_per_second": 0.738
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 4.0,
+    "eval_accuracy": 0.9076675735839723,
+    "eval_loss": 0.3249448239803314,
+    "eval_runtime": 126.4373,
+    "eval_samples": 40430,
+    "eval_samples_per_second": 319.763,
+    "eval_steps_per_second": 4.999
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 4.0,
+    "train_loss": 0.17184808974027801,
+    "train_runtime": 15393.9025,
+    "train_samples": 363846,
+    "train_samples_per_second": 94.543,
+    "train_steps_per_second": 0.738
+}

trainer_state.json

@@ -0,0 +1,193 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.9989446838448686,
+  "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": 4.780084447572132e-05,
+      "loss": 0.4112,
+      "step": 500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.560168895144265e-05,
+      "loss": 0.33,
+      "step": 1000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.3402533427163974e-05,
+      "loss": 0.2981,
+      "step": 1500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.120337790288529e-05,
+      "loss": 0.2886,
+      "step": 2000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.900422237860662e-05,
+      "loss": 0.2764,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8866930497155577,
+      "eval_loss": 0.26510927081108093,
+      "eval_runtime": 127.0864,
+      "eval_samples_per_second": 318.13,
+      "eval_steps_per_second": 4.973,
+      "step": 2842
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.680506685432794e-05,
+      "loss": 0.24,
+      "step": 3000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.461030964109782e-05,
+      "loss": 0.192,
+      "step": 3500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.2411154116819145e-05,
+      "loss": 0.1979,
+      "step": 4000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.0211998592540464e-05,
+      "loss": 0.193,
+      "step": 4500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 2.801284306826179e-05,
+      "loss": 0.1889,
+      "step": 5000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.5813687543983113e-05,
+      "loss": 0.1888,
+      "step": 5500
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9001978728666832,
+      "eval_loss": 0.2454005628824234,
+      "eval_runtime": 126.6764,
+      "eval_samples_per_second": 319.16,
+      "eval_steps_per_second": 4.989,
+      "step": 5685
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.3618930330752993e-05,
+      "loss": 0.1407,
+      "step": 6000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.1419774806474316e-05,
+      "loss": 0.115,
+      "step": 6500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.9220619282195635e-05,
+      "loss": 0.1184,
+      "step": 7000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.702146375791696e-05,
+      "loss": 0.1144,
+      "step": 7500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.4822308233638283e-05,
+      "loss": 0.1092,
+      "step": 8000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.2623152709359607e-05,
+      "loss": 0.1161,
+      "step": 8500
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9061835270838486,
+      "eval_loss": 0.2765159606933594,
+      "eval_runtime": 127.1372,
+      "eval_samples_per_second": 318.003,
+      "eval_steps_per_second": 4.971,
+      "step": 8528
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.0428395496129487e-05,
+      "loss": 0.0701,
+      "step": 9000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 8.22923997185081e-06,
+      "loss": 0.069,
+      "step": 9500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.030084447572133e-06,
+      "loss": 0.0678,
+      "step": 10000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.8309289232934554e-06,
+      "loss": 0.0668,
+      "step": 10500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.6317733990147786e-06,
+      "loss": 0.0663,
+      "step": 11000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9076675735839723,
+      "eval_loss": 0.3249448239803314,
+      "eval_runtime": 126.6579,
+      "eval_samples_per_second": 319.206,
+      "eval_steps_per_second": 4.99,
+      "step": 11368
+    },
+    {
+      "epoch": 4.0,
+      "step": 11368,
+      "total_flos": 3.8283342643623936e+17,
+      "train_loss": 0.17184808974027801,
+      "train_runtime": 15393.9025,
+      "train_samples_per_second": 94.543,
+      "train_steps_per_second": 0.738
+    }
+  ],
+  "max_steps": 11368,
+  "num_train_epochs": 4,
+  "total_flos": 3.8283342643623936e+17,
+  "trial_name": null,
+  "trial_params": null
+}