End of training

all_results.json

@@ -1,14 +1,14 @@
 {
-    "epoch": 3.0,
-    "eval_loss": 1.920763373374939,
-    "eval_runtime": 80.7765,
+    "epoch": 8.0,
+    "eval_loss": 1.946905255317688,
+    "eval_runtime": 80.8062,
     "eval_samples": 5193,
-    "eval_samples_per_second": 64.289,
-    "eval_steps_per_second": 8.047,
-    "perplexity": 6.826167395302532,
-    "train_loss": 2.110191111022448,
-    "train_runtime": 17461.8774,
+    "eval_samples_per_second": 64.265,
+    "eval_steps_per_second": 8.044,
+    "perplexity": 7.006969210814188,
+    "train_loss": 2.007555020303887,
+    "train_runtime": 46638.2691,
     "train_samples": 98673,
-    "train_samples_per_second": 16.952,
+    "train_samples_per_second": 16.926,
     "train_steps_per_second": 0.265
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.0,
-    "eval_loss": 1.920763373374939,
-    "eval_runtime": 80.7765,
+    "epoch": 8.0,
+    "eval_loss": 1.946905255317688,
+    "eval_runtime": 80.8062,
     "eval_samples": 5193,
-    "eval_samples_per_second": 64.289,
-    "eval_steps_per_second": 8.047,
-    "perplexity": 6.826167395302532
+    "eval_samples_per_second": 64.265,
+    "eval_steps_per_second": 8.044,
+    "perplexity": 7.006969210814188
 }

runs/Dec03_09-11-10_jbuhmann/events.out.tfevents.1638522692.jbuhmann.15391.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ba3d01ba8737adf4a184071e0dc40283d0d880d753d11d771842000162cb3c74
-size 9083
+oid sha256:56c406bf2bffd5610c48cd40266d4b62cfcf02154f0b755c2c8030e9c76cb2bd
+size 9437

runs/Dec03_09-11-10_jbuhmann/events.out.tfevents.1638569412.jbuhmann.15391.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a090e2d4deb7ea2c1721854597cad6e2019b6bb4d9913f1063fac5cd207847c1
+size 311

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "train_loss": 2.110191111022448,
-    "train_runtime": 17461.8774,
+    "epoch": 8.0,
+    "train_loss": 2.007555020303887,
+    "train_runtime": 46638.2691,
     "train_samples": 98673,
-    "train_samples_per_second": 16.952,
+    "train_samples_per_second": 16.926,
     "train_steps_per_second": 0.265
 }

trainer_state.json

@@ -1,103 +1,233 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.0,
-  "global_step": 4626,
+  "epoch": 8.0,
+  "global_step": 12336,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.32,
-      "learning_rate": 1.7838305231301342e-05,
-      "loss": 2.2294,
+      "learning_rate": 1.9189364461738006e-05,
+      "loss": 2.2297,
       "step": 500
     },
     {
       "epoch": 0.65,
-      "learning_rate": 1.5676610462602683e-05,
-      "loss": 2.1683,
+      "learning_rate": 1.8378728923476004e-05,
+      "loss": 2.1699,
       "step": 1000
     },
     {
       "epoch": 0.97,
-      "learning_rate": 1.3514915693904023e-05,
-      "loss": 2.1713,
+      "learning_rate": 1.756809338521401e-05,
+      "loss": 2.1739,
       "step": 1500
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.009819507598877,
-      "eval_runtime": 81.3789,
-      "eval_samples_per_second": 63.813,
-      "eval_steps_per_second": 7.987,
+      "eval_loss": 2.015004873275757,
+      "eval_runtime": 81.2342,
+      "eval_samples_per_second": 63.926,
+      "eval_steps_per_second": 8.002,
       "step": 1542
     },
     {
       "epoch": 1.3,
-      "learning_rate": 1.1353220925205362e-05,
-      "loss": 2.1142,
+      "learning_rate": 1.675745784695201e-05,
+      "loss": 2.1159,
       "step": 2000
     },
     {
       "epoch": 1.62,
-      "learning_rate": 9.191526156506702e-06,
-      "loss": 2.0871,
+      "learning_rate": 1.5946822308690015e-05,
+      "loss": 2.0903,
       "step": 2500
     },
     {
       "epoch": 1.95,
-      "learning_rate": 7.029831387808041e-06,
-      "loss": 2.0736,
+      "learning_rate": 1.5136186770428017e-05,
+      "loss": 2.0759,
       "step": 3000
     },
     {
       "epoch": 2.0,
-      "eval_loss": 1.9852919578552246,
-      "eval_runtime": 81.4601,
-      "eval_samples_per_second": 63.749,
-      "eval_steps_per_second": 7.979,
+      "eval_loss": 1.9917508363723755,
+      "eval_runtime": 81.2881,
+      "eval_samples_per_second": 63.884,
+      "eval_steps_per_second": 7.996,
       "step": 3084
     },
     {
       "epoch": 2.27,
-      "learning_rate": 4.8681366191093824e-06,
-      "loss": 2.0674,
+      "learning_rate": 1.432555123216602e-05,
+      "loss": 2.0628,
       "step": 3500
     },
     {
       "epoch": 2.59,
-      "learning_rate": 2.706441850410722e-06,
-      "loss": 2.0482,
+      "learning_rate": 1.3514915693904023e-05,
+      "loss": 2.0423,
       "step": 4000
     },
     {
       "epoch": 2.92,
-      "learning_rate": 5.447470817120623e-07,
-      "loss": 2.0543,
+      "learning_rate": 1.2704280155642024e-05,
+      "loss": 2.0453,
       "step": 4500
     },
     {
       "epoch": 3.0,
-      "eval_loss": 2.013369083404541,
-      "eval_runtime": 81.3121,
-      "eval_samples_per_second": 63.865,
-      "eval_steps_per_second": 7.994,
+      "eval_loss": 2.0132408142089844,
+      "eval_runtime": 81.2855,
+      "eval_samples_per_second": 63.886,
+      "eval_steps_per_second": 7.997,
       "step": 4626
     },
     {
-      "epoch": 3.0,
-      "step": 4626,
-      "total_flos": 7.791321717294797e+16,
-      "train_loss": 2.110191111022448,
-      "train_runtime": 17461.8774,
-      "train_samples_per_second": 16.952,
+      "epoch": 3.24,
+      "learning_rate": 1.1893644617380028e-05,
+      "loss": 2.0054,
+      "step": 5000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.108300907911803e-05,
+      "loss": 1.9898,
+      "step": 5500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.027237354085603e-05,
+      "loss": 1.9936,
+      "step": 6000
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 1.9341049194335938,
+      "eval_runtime": 81.6406,
+      "eval_samples_per_second": 63.608,
+      "eval_steps_per_second": 7.962,
+      "step": 6168
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 9.461738002594033e-06,
+      "loss": 1.99,
+      "step": 6500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 8.651102464332037e-06,
+      "loss": 1.9735,
+      "step": 7000
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 7.84046692607004e-06,
+      "loss": 1.9659,
+      "step": 7500
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 1.9140448570251465,
+      "eval_runtime": 81.5402,
+      "eval_samples_per_second": 63.686,
+      "eval_steps_per_second": 7.972,
+      "step": 7710
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 7.029831387808041e-06,
+      "loss": 1.9519,
+      "step": 8000
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 6.219195849546045e-06,
+      "loss": 1.933,
+      "step": 8500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 5.4085603112840465e-06,
+      "loss": 1.9545,
+      "step": 9000
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 1.941838026046753,
+      "eval_runtime": 81.6484,
+      "eval_samples_per_second": 63.602,
+      "eval_steps_per_second": 7.961,
+      "step": 9252
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 4.5979247730220496e-06,
+      "loss": 1.9394,
+      "step": 9500
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 3.7872892347600522e-06,
+      "loss": 1.9197,
+      "step": 10000
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 2.9766536964980545e-06,
+      "loss": 1.9104,
+      "step": 10500
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 1.9178550243377686,
+      "eval_runtime": 81.3195,
+      "eval_samples_per_second": 63.859,
+      "eval_steps_per_second": 7.993,
+      "step": 10794
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 2.166018158236057e-06,
+      "loss": 1.909,
+      "step": 11000
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.3553826199740596e-06,
+      "loss": 1.8942,
+      "step": 11500
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 5.447470817120623e-07,
+      "loss": 1.8991,
+      "step": 12000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 1.9157277345657349,
+      "eval_runtime": 81.5335,
+      "eval_samples_per_second": 63.692,
+      "eval_steps_per_second": 7.972,
+      "step": 12336
+    },
+    {
+      "epoch": 8.0,
+      "step": 12336,
+      "total_flos": 2.0776857912786125e+17,
+      "train_loss": 2.007555020303887,
+      "train_runtime": 46638.2691,
+      "train_samples_per_second": 16.926,
       "train_steps_per_second": 0.265
     }
   ],
-  "max_steps": 4626,
-  "num_train_epochs": 3,
-  "total_flos": 7.791321717294797e+16,
+  "max_steps": 12336,
+  "num_train_epochs": 8,
+  "total_flos": 2.0776857912786125e+17,
   "trial_name": null,
   "trial_params": null
 }