End of training

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 8.0,
-    "eval_loss": 2.0533692836761475,
-    "eval_runtime": 80.7384,
+    "eval_loss": 1.9716644287109375,
+    "eval_runtime": 80.7808,
     "eval_samples": 5193,
-    "eval_samples_per_second": 64.319,
-    "eval_steps_per_second": 8.051,
-    "perplexity": 7.794117509150382,
-    "train_loss": 2.459805450241544,
-    "train_runtime": 46512.8854,
+    "eval_samples_per_second": 64.285,
+    "eval_steps_per_second": 8.046,
+    "perplexity": 7.18262150622839,
+    "train_loss": 2.136193385538579,
+    "train_runtime": 46556.9069,
     "train_samples": 98673,
-    "train_samples_per_second": 16.971,
+    "train_samples_per_second": 16.955,
     "train_steps_per_second": 0.265
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 8.0,
-    "eval_loss": 2.0533692836761475,
-    "eval_runtime": 80.7384,
+    "eval_loss": 1.9716644287109375,
+    "eval_runtime": 80.7808,
     "eval_samples": 5193,
-    "eval_samples_per_second": 64.319,
-    "eval_steps_per_second": 8.051,
-    "perplexity": 7.794117509150382
+    "eval_samples_per_second": 64.285,
+    "eval_steps_per_second": 8.046,
+    "perplexity": 7.18262150622839
 }

runs/Dec21_13-49-08_jbuhmann/events.out.tfevents.1640094568.jbuhmann.7947.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db4ed3d99a342d393b6d1cb5e24f1a5c4f8b2fb558dc85ef593dbb176e00717d
-size 9119
+oid sha256:141ca83b6acbfc4954e9b74f24bb41738b901790a0d1a7c545a4d5ead86638e9
+size 9473

runs/Dec21_13-49-08_jbuhmann/events.out.tfevents.1640141207.jbuhmann.7947.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c1b258f7c0d45ac894e84b7acfcdb1f8dd016b1513e3351bb988bf376a2136f
+size 311

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 8.0,
-    "train_loss": 2.459805450241544,
-    "train_runtime": 46512.8854,
+    "train_loss": 2.136193385538579,
+    "train_runtime": 46556.9069,
     "train_samples": 98673,
-    "train_samples_per_second": 16.971,
+    "train_samples_per_second": 16.955,
     "train_steps_per_second": 0.265
 }

trainer_state.json

@@ -10,218 +10,218 @@
     {
       "epoch": 0.32,
       "learning_rate": 1.9189364461738006e-05,
-      "loss": 2.8003,
+      "loss": 2.3643,
       "step": 500
     },
     {
       "epoch": 0.65,
       "learning_rate": 1.8378728923476004e-05,
-      "loss": 2.6656,
+      "loss": 2.2958,
       "step": 1000
     },
     {
       "epoch": 0.97,
       "learning_rate": 1.756809338521401e-05,
-      "loss": 2.6501,
+      "loss": 2.2954,
       "step": 1500
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.1010942459106445,
-      "eval_runtime": 81.2019,
-      "eval_samples_per_second": 63.952,
-      "eval_steps_per_second": 8.005,
+      "eval_loss": 2.0372068881988525,
+      "eval_runtime": 81.4002,
+      "eval_samples_per_second": 63.796,
+      "eval_steps_per_second": 7.985,
       "step": 1542
     },
     {
       "epoch": 1.3,
       "learning_rate": 1.675745784695201e-05,
-      "loss": 2.5775,
+      "loss": 2.2408,
       "step": 2000
     },
     {
       "epoch": 1.62,
       "learning_rate": 1.5946822308690015e-05,
-      "loss": 2.545,
+      "loss": 2.213,
       "step": 2500
     },
     {
       "epoch": 1.95,
       "learning_rate": 1.5136186770428017e-05,
-      "loss": 2.5183,
+      "loss": 2.2015,
       "step": 3000
     },
     {
       "epoch": 2.0,
-      "eval_loss": 2.085500478744507,
-      "eval_runtime": 81.2927,
-      "eval_samples_per_second": 63.88,
-      "eval_steps_per_second": 7.996,
+      "eval_loss": 2.0103893280029297,
+      "eval_runtime": 81.4187,
+      "eval_samples_per_second": 63.781,
+      "eval_steps_per_second": 7.983,
       "step": 3084
     },
     {
       "epoch": 2.27,
       "learning_rate": 1.432555123216602e-05,
-      "loss": 2.5098,
+      "loss": 2.1898,
       "step": 3500
     },
     {
       "epoch": 2.59,
       "learning_rate": 1.3514915693904023e-05,
-      "loss": 2.4869,
+      "loss": 2.1665,
       "step": 4000
     },
     {
       "epoch": 2.92,
       "learning_rate": 1.2704280155642024e-05,
-      "loss": 2.4835,
+      "loss": 2.1661,
       "step": 4500
     },
     {
       "epoch": 3.0,
-      "eval_loss": 2.1157290935516357,
-      "eval_runtime": 81.2608,
-      "eval_samples_per_second": 63.905,
-      "eval_steps_per_second": 7.999,
+      "eval_loss": 2.0372352600097656,
+      "eval_runtime": 81.1974,
+      "eval_samples_per_second": 63.955,
+      "eval_steps_per_second": 8.005,
       "step": 4626
     },
     {
       "epoch": 3.24,
       "learning_rate": 1.1893644617380028e-05,
-      "loss": 2.4413,
+      "loss": 2.1286,
       "step": 5000
     },
     {
       "epoch": 3.57,
       "learning_rate": 1.108300907911803e-05,
-      "loss": 2.4289,
+      "loss": 2.1179,
       "step": 5500
     },
     {
       "epoch": 3.89,
       "learning_rate": 1.027237354085603e-05,
-      "loss": 2.4372,
+      "loss": 2.1186,
       "step": 6000
     },
     {
       "epoch": 4.0,
-      "eval_loss": 2.0268728733062744,
-      "eval_runtime": 81.2557,
-      "eval_samples_per_second": 63.909,
-      "eval_steps_per_second": 7.999,
+      "eval_loss": 1.9548556804656982,
+      "eval_runtime": 81.2263,
+      "eval_samples_per_second": 63.932,
+      "eval_steps_per_second": 8.002,
       "step": 6168
     },
     {
       "epoch": 4.22,
       "learning_rate": 9.461738002594033e-06,
-      "loss": 2.4307,
+      "loss": 2.1206,
       "step": 6500
     },
     {
       "epoch": 4.54,
       "learning_rate": 8.651102464332037e-06,
-      "loss": 2.4153,
+      "loss": 2.1017,
       "step": 7000
     },
     {
       "epoch": 4.86,
       "learning_rate": 7.84046692607004e-06,
-      "loss": 2.4087,
+      "loss": 2.0939,
       "step": 7500
     },
     {
       "epoch": 5.0,
-      "eval_loss": 2.0193827152252197,
-      "eval_runtime": 81.2425,
-      "eval_samples_per_second": 63.92,
-      "eval_steps_per_second": 8.001,
+      "eval_loss": 1.9438201189041138,
+      "eval_runtime": 81.1944,
+      "eval_samples_per_second": 63.958,
+      "eval_steps_per_second": 8.005,
       "step": 7710
     },
     {
       "epoch": 5.19,
       "learning_rate": 7.029831387808041e-06,
-      "loss": 2.393,
+      "loss": 2.084,
       "step": 8000
     },
     {
       "epoch": 5.51,
       "learning_rate": 6.219195849546045e-06,
-      "loss": 2.3759,
+      "loss": 2.0629,
       "step": 8500
     },
     {
       "epoch": 5.84,
       "learning_rate": 5.4085603112840465e-06,
-      "loss": 2.394,
+      "loss": 2.0867,
       "step": 9000
     },
     {
       "epoch": 6.0,
-      "eval_loss": 2.0468995571136475,
-      "eval_runtime": 81.2739,
-      "eval_samples_per_second": 63.895,
-      "eval_steps_per_second": 7.998,
+      "eval_loss": 1.9647932052612305,
+      "eval_runtime": 81.2286,
+      "eval_samples_per_second": 63.931,
+      "eval_steps_per_second": 8.002,
       "step": 9252
     },
     {
       "epoch": 6.16,
       "learning_rate": 4.5979247730220496e-06,
-      "loss": 2.3816,
+      "loss": 2.0697,
       "step": 9500
     },
     {
       "epoch": 6.49,
       "learning_rate": 3.7872892347600522e-06,
-      "loss": 2.3603,
+      "loss": 2.0516,
       "step": 10000
     },
     {
       "epoch": 6.81,
       "learning_rate": 2.9766536964980545e-06,
-      "loss": 2.3576,
+      "loss": 2.0462,
       "step": 10500
     },
     {
       "epoch": 7.0,
-      "eval_loss": 2.029569625854492,
-      "eval_runtime": 81.1582,
-      "eval_samples_per_second": 63.986,
-      "eval_steps_per_second": 8.009,
+      "eval_loss": 1.9464805126190186,
+      "eval_runtime": 81.3559,
+      "eval_samples_per_second": 63.831,
+      "eval_steps_per_second": 7.99,
       "step": 10794
     },
     {
       "epoch": 7.13,
       "learning_rate": 2.166018158236057e-06,
-      "loss": 2.3497,
+      "loss": 2.0434,
       "step": 11000
     },
     {
       "epoch": 7.46,
       "learning_rate": 1.3553826199740596e-06,
-      "loss": 2.3399,
+      "loss": 2.0279,
       "step": 11500
     },
     {
       "epoch": 7.78,
       "learning_rate": 5.447470817120623e-07,
-      "loss": 2.34,
+      "loss": 2.0315,
       "step": 12000
     },
     {
       "epoch": 8.0,
-      "eval_loss": 2.0294876098632812,
-      "eval_runtime": 81.1819,
-      "eval_samples_per_second": 63.967,
-      "eval_steps_per_second": 8.007,
+      "eval_loss": 1.9411877393722534,
+      "eval_runtime": 81.3329,
+      "eval_samples_per_second": 63.849,
+      "eval_steps_per_second": 7.992,
       "step": 12336
     },
     {
       "epoch": 8.0,
       "step": 12336,
       "total_flos": 2.0776857912786125e+17,
-      "train_loss": 2.459805450241544,
-      "train_runtime": 46512.8854,
-      "train_samples_per_second": 16.971,
+      "train_loss": 2.136193385538579,
+      "train_runtime": 46556.9069,
+      "train_samples_per_second": 16.955,
       "train_steps_per_second": 0.265
     }
   ],