End of training

README.md

@@ -2,7 +2,10 @@
 license: apache-2.0
 base_model: bert-base-uncased
 tags:
+- question-answering
 - generated_from_trainer
+datasets:
+- squad
 model-index:
 - name: out
   results: []
@@ -13,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # out
 
-This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset.
+This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the squad dataset.
 
 ## Model description
 

all_results.json

@@ -1,11 +1,11 @@
 {
-    "epoch": 50.0,
+    "epoch": 100.0,
     "eval_samples": 3000,
-    "exact_match": 55.833333333333336,
-    "f1": 65.67795577455873,
-    "train_loss": 0.1556171702324076,
-    "train_runtime": 1356.8935,
+    "exact_match": 48.3,
+    "f1": 59.76387313784877,
+    "train_loss": 0.1459447158651149,
+    "train_runtime": 2721.3121,
     "train_samples": 3000,
-    "train_samples_per_second": 110.547,
-    "train_steps_per_second": 3.464
+    "train_samples_per_second": 110.241,
+    "train_steps_per_second": 3.454
 }

args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a6483b4338266a431a726740f9d0603e8af5b5ea355462122d055eb126bb9e1e
+oid sha256:0f140c580ce20b0c341613069c779fdc1ae1caec64fbe449d785823651e822ba
 size 5652

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:630b581cec250940167a383435d09347c398cba0c12a4e660ef2760f5d1b9877
-size 14106174
+oid sha256:4e93beefb2d2ba12031615060145a1efb9f0ba9b4e4b3b542feda79321cf90b3
+size 14014825

eval_results.json

@@ -1,6 +1,6 @@
 {
-    "epoch": 50.0,
+    "epoch": 100.0,
     "eval_samples": 3000,
-    "exact_match": 55.833333333333336,
-    "f1": 65.67795577455873
+    "exact_match": 48.3,
+    "f1": 59.76387313784877
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
-    "train_loss": 0.1556171702324076,
-    "train_runtime": 1356.8935,
+    "epoch": 100.0,
+    "train_loss": 0.1459447158651149,
+    "train_runtime": 2721.3121,
     "train_samples": 3000,
-    "train_samples_per_second": 110.547,
-    "train_steps_per_second": 3.464
+    "train_samples_per_second": 110.241,
+    "train_steps_per_second": 3.454
 }

trainer_state.json

@@ -1,82 +1,136 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 50.0,
+  "epoch": 100.0,
   "eval_steps": 500,
-  "global_step": 4700,
+  "global_step": 9400,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 5.32,
-      "learning_rate": 4.468085106382979e-05,
-      "loss": 1.1017,
+      "learning_rate": 4.734042553191489e-05,
+      "loss": 1.0988,
       "step": 500
     },
     {
       "epoch": 10.64,
-      "learning_rate": 3.936170212765958e-05,
-      "loss": 0.1281,
+      "learning_rate": 4.468085106382979e-05,
+      "loss": 0.1377,
       "step": 1000
     },
     {
       "epoch": 15.96,
-      "learning_rate": 3.4042553191489365e-05,
-      "loss": 0.0555,
+      "learning_rate": 4.2021276595744684e-05,
+      "loss": 0.0645,
       "step": 1500
     },
     {
       "epoch": 21.28,
-      "learning_rate": 2.8723404255319154e-05,
-      "loss": 0.027,
+      "learning_rate": 3.936170212765958e-05,
+      "loss": 0.0425,
       "step": 2000
     },
     {
       "epoch": 26.6,
-      "learning_rate": 2.340425531914894e-05,
-      "loss": 0.0195,
+      "learning_rate": 3.670212765957447e-05,
+      "loss": 0.0245,
       "step": 2500
     },
     {
       "epoch": 31.91,
-      "learning_rate": 1.8085106382978724e-05,
-      "loss": 0.0135,
+      "learning_rate": 3.4042553191489365e-05,
+      "loss": 0.0228,
       "step": 3000
     },
     {
       "epoch": 37.23,
-      "learning_rate": 1.2765957446808511e-05,
-      "loss": 0.0103,
+      "learning_rate": 3.1382978723404254e-05,
+      "loss": 0.0251,
       "step": 3500
     },
     {
       "epoch": 42.55,
-      "learning_rate": 7.446808510638298e-06,
-      "loss": 0.0173,
+      "learning_rate": 2.8723404255319154e-05,
+      "loss": 0.0384,
       "step": 4000
     },
     {
       "epoch": 47.87,
-      "learning_rate": 2.1276595744680853e-06,
-      "loss": 0.0432,
+      "learning_rate": 2.6063829787234046e-05,
+      "loss": 0.0429,
       "step": 4500
     },
     {
-      "epoch": 50.0,
-      "step": 4700,
-      "total_flos": 2.93958851328e+16,
-      "train_loss": 0.1556171702324076,
-      "train_runtime": 1356.8935,
-      "train_samples_per_second": 110.547,
-      "train_steps_per_second": 3.464
+      "epoch": 53.19,
+      "learning_rate": 2.340425531914894e-05,
+      "loss": 0.055,
+      "step": 5000
+    },
+    {
+      "epoch": 58.51,
+      "learning_rate": 2.074468085106383e-05,
+      "loss": 0.0883,
+      "step": 5500
+    },
+    {
+      "epoch": 63.83,
+      "learning_rate": 1.8085106382978724e-05,
+      "loss": 0.1385,
+      "step": 6000
+    },
+    {
+      "epoch": 69.15,
+      "learning_rate": 1.5425531914893617e-05,
+      "loss": 0.2111,
+      "step": 6500
+    },
+    {
+      "epoch": 74.47,
+      "learning_rate": 1.2765957446808511e-05,
+      "loss": 0.2821,
+      "step": 7000
+    },
+    {
+      "epoch": 79.79,
+      "learning_rate": 1.0106382978723404e-05,
+      "loss": 0.1479,
+      "step": 7500
+    },
+    {
+      "epoch": 85.11,
+      "learning_rate": 7.446808510638298e-06,
+      "loss": 0.1729,
+      "step": 8000
+    },
+    {
+      "epoch": 90.43,
+      "learning_rate": 4.787234042553191e-06,
+      "loss": 0.0951,
+      "step": 8500
+    },
+    {
+      "epoch": 95.74,
+      "learning_rate": 2.1276595744680853e-06,
+      "loss": 0.0391,
+      "step": 9000
+    },
+    {
+      "epoch": 100.0,
+      "step": 9400,
+      "total_flos": 5.87917702656e+16,
+      "train_loss": 0.1459447158651149,
+      "train_runtime": 2721.3121,
+      "train_samples_per_second": 110.241,
+      "train_steps_per_second": 3.454
     }
   ],
   "logging_steps": 500,
-  "max_steps": 4700,
-  "num_train_epochs": 50,
+  "max_steps": 9400,
+  "num_train_epochs": 100,
   "save_steps": 500,
-  "total_flos": 2.93958851328e+16,
+  "total_flos": 5.87917702656e+16,
   "trial_name": null,
   "trial_params": null
 }