End of training

all_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 2.0,
-    "eval_accuracy": 0.7105720509450737,
-    "eval_loss": 1.2192070484161377,
-    "eval_runtime": 4409.7065,
+    "eval_accuracy": 0.7436041304518132,
+    "eval_loss": 1.0827183723449707,
+    "eval_runtime": 4164.5239,
     "eval_samples": 134942,
-    "eval_samples_per_second": 30.601,
-    "eval_steps_per_second": 7.65,
-    "perplexity": 3.384502922346475,
-    "train_loss": 1.4528967095198742,
-    "train_runtime": 23861.4316,
+    "eval_samples_per_second": 32.403,
+    "eval_steps_per_second": 3.24,
+    "perplexity": 2.9526951763129543,
+    "train_loss": 0.05643905973204775,
+    "train_runtime": 943.9109,
     "train_samples": 134942,
-    "train_samples_per_second": 11.31,
-    "train_steps_per_second": 1.131
+    "train_samples_per_second": 285.921,
+    "train_steps_per_second": 28.594
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.7105720509450737,
-    "eval_loss": 1.2192070484161377,
-    "eval_runtime": 4409.7065,
+    "epoch": 2.0,
+    "eval_accuracy": 0.7436041304518132,
+    "eval_loss": 1.0827183723449707,
+    "eval_runtime": 4164.5239,
     "eval_samples": 134942,
-    "eval_samples_per_second": 30.601,
-    "eval_steps_per_second": 7.65,
-    "perplexity": 3.384502922346475
+    "eval_samples_per_second": 32.403,
+    "eval_steps_per_second": 3.24,
+    "perplexity": 2.9526951763129543
 }

runs/Aug19_01-34-35_ac5e0a9e9d4e/events.out.tfevents.1660878687.ac5e0a9e9d4e.2306.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4272eecad0977928c92af9b3050ac1093aae3e57871407df9ac975512f8a30a
+size 369

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.0,
-    "train_loss": 1.4528967095198742,
-    "train_runtime": 23861.4316,
+    "train_loss": 0.05643905973204775,
+    "train_runtime": 943.9109,
     "train_samples": 134942,
-    "train_samples_per_second": 11.31,
-    "train_steps_per_second": 1.131
+    "train_samples_per_second": 285.921,
+    "train_steps_per_second": 28.594
 }

trainer_state.json

@@ -329,10 +329,10 @@
       "epoch": 2.0,
       "step": 26990,
       "total_flos": 2.5064145824199475e+17,
-      "train_loss": 1.4528967095198742,
-      "train_runtime": 23861.4316,
-      "train_samples_per_second": 11.31,
-      "train_steps_per_second": 1.131
+      "train_loss": 0.05643905973204775,
+      "train_runtime": 943.9109,
+      "train_samples_per_second": 285.921,
+      "train_steps_per_second": 28.594
     }
   ],
   "max_steps": 26990,