End of training

README.md

@@ -2,7 +2,10 @@
 license: apache-2.0
 base_model: distilbert-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 [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
+This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the squad dataset.
 
 ## Model description
 

all_results.json

@@ -1,11 +1,11 @@
 {
-    "epoch": 50.0,
+    "epoch": 100.0,
     "eval_samples": 5000,
-    "exact_match": 26.0,
-    "f1": 38.36451648734593,
-    "train_loss": 0.5251886637499378,
-    "train_runtime": 1204.0321,
+    "exact_match": 38.78,
+    "f1": 49.77281879399985,
+    "train_loss": 0.15928955945619352,
+    "train_runtime": 2364.7944,
     "train_samples": 5000,
-    "train_samples_per_second": 207.636,
-    "train_steps_per_second": 6.52
+    "train_samples_per_second": 211.435,
+    "train_steps_per_second": 6.639
 }

args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9bc713c84ef398187cd03f1d0ce712589fcfdb11714e0b3ed833239e8567fa85
+oid sha256:e487e45369f7ed4c3a457a7a1bc11000c203d2ec9bcad825bd40329736fe5301
 size 5652

eval_nbest_predictions.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a4ca005506c01dfbfd7147903614c6b992f9189f7dc07f7befbae177fb7bd41
-size 23339407
+oid sha256:30817220b7512e944ac3582c72535a287e4394685838e42fa00f5fba592aef2e
+size 23123501

eval_results.json

@@ -1,6 +1,6 @@
 {
-    "epoch": 50.0,
+    "epoch": 100.0,
     "eval_samples": 5000,
-    "exact_match": 26.0,
-    "f1": 38.36451648734593
+    "exact_match": 38.78,
+    "f1": 49.77281879399985
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
-    "train_loss": 0.5251886637499378,
-    "train_runtime": 1204.0321,
+    "epoch": 100.0,
+    "train_loss": 0.15928955945619352,
+    "train_runtime": 2364.7944,
     "train_samples": 5000,
-    "train_samples_per_second": 207.636,
-    "train_steps_per_second": 6.52
+    "train_samples_per_second": 211.435,
+    "train_steps_per_second": 6.639
 }

trainer_state.json

@@ -1,118 +1,214 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 50.0,
+  "epoch": 100.0,
   "eval_steps": 500,
-  "global_step": 7850,
+  "global_step": 15700,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 3.18,
-      "learning_rate": 4.681528662420383e-05,
-      "loss": 1.5828,
+      "learning_rate": 4.840764331210191e-05,
+      "loss": 1.5811,
       "step": 500
     },
     {
       "epoch": 6.37,
-      "learning_rate": 4.3630573248407646e-05,
-      "loss": 0.2954,
+      "learning_rate": 4.681528662420383e-05,
+      "loss": 0.2979,
       "step": 1000
     },
     {
       "epoch": 9.55,
-      "learning_rate": 4.044585987261147e-05,
-      "loss": 0.123,
+      "learning_rate": 4.522292993630574e-05,
+      "loss": 0.1303,
       "step": 1500
     },
     {
       "epoch": 12.74,
-      "learning_rate": 3.7261146496815283e-05,
-      "loss": 0.078,
+      "learning_rate": 4.3630573248407646e-05,
+      "loss": 0.0811,
       "step": 2000
     },
     {
       "epoch": 15.92,
-      "learning_rate": 3.407643312101911e-05,
-      "loss": 0.0562,
+      "learning_rate": 4.2038216560509556e-05,
+      "loss": 0.0603,
       "step": 2500
     },
     {
       "epoch": 19.11,
-      "learning_rate": 3.089171974522293e-05,
-      "loss": 0.0447,
+      "learning_rate": 4.044585987261147e-05,
+      "loss": 0.0594,
       "step": 3000
     },
     {
       "epoch": 22.29,
-      "learning_rate": 2.7707006369426753e-05,
-      "loss": 0.053,
+      "learning_rate": 3.885350318471338e-05,
+      "loss": 0.0659,
       "step": 3500
     },
     {
       "epoch": 25.48,
-      "learning_rate": 2.4522292993630575e-05,
-      "loss": 0.0765,
+      "learning_rate": 3.7261146496815283e-05,
+      "loss": 0.0848,
       "step": 4000
     },
     {
       "epoch": 28.66,
-      "learning_rate": 2.1337579617834397e-05,
-      "loss": 0.1098,
+      "learning_rate": 3.56687898089172e-05,
+      "loss": 0.1071,
       "step": 4500
     },
     {
       "epoch": 31.85,
-      "learning_rate": 1.8152866242038215e-05,
-      "loss": 0.2508,
+      "learning_rate": 3.407643312101911e-05,
+      "loss": 0.186,
       "step": 5000
     },
     {
       "epoch": 35.03,
-      "learning_rate": 1.4968152866242039e-05,
-      "loss": 0.516,
+      "learning_rate": 3.248407643312102e-05,
+      "loss": 0.2398,
       "step": 5500
     },
     {
       "epoch": 38.22,
-      "learning_rate": 1.178343949044586e-05,
-      "loss": 0.8691,
+      "learning_rate": 3.089171974522293e-05,
+      "loss": 0.3481,
       "step": 6000
     },
     {
       "epoch": 41.4,
-      "learning_rate": 8.598726114649681e-06,
-      "loss": 0.9585,
+      "learning_rate": 2.929936305732484e-05,
+      "loss": 0.4227,
       "step": 6500
     },
     {
       "epoch": 44.59,
-      "learning_rate": 5.414012738853504e-06,
-      "loss": 1.1375,
+      "learning_rate": 2.7707006369426753e-05,
+      "loss": 0.4091,
       "step": 7000
     },
     {
       "epoch": 47.77,
-      "learning_rate": 2.229299363057325e-06,
-      "loss": 1.2299,
+      "learning_rate": 2.6114649681528662e-05,
+      "loss": 0.3251,
       "step": 7500
     },
     {
-      "epoch": 50.0,
-      "step": 7850,
-      "total_flos": 2.4497456256e+16,
-      "train_loss": 0.5251886637499378,
-      "train_runtime": 1204.0321,
-      "train_samples_per_second": 207.636,
-      "train_steps_per_second": 6.52
+      "epoch": 50.96,
+      "learning_rate": 2.4522292993630575e-05,
+      "loss": 0.2278,
+      "step": 8000
+    },
+    {
+      "epoch": 54.14,
+      "learning_rate": 2.2929936305732484e-05,
+      "loss": 0.1192,
+      "step": 8500
+    },
+    {
+      "epoch": 57.32,
+      "learning_rate": 2.1337579617834397e-05,
+      "loss": 0.0682,
+      "step": 9000
+    },
+    {
+      "epoch": 60.51,
+      "learning_rate": 1.974522292993631e-05,
+      "loss": 0.0419,
+      "step": 9500
+    },
+    {
+      "epoch": 63.69,
+      "learning_rate": 1.8152866242038215e-05,
+      "loss": 0.0289,
+      "step": 10000
+    },
+    {
+      "epoch": 66.88,
+      "learning_rate": 1.6560509554140128e-05,
+      "loss": 0.0205,
+      "step": 10500
+    },
+    {
+      "epoch": 70.06,
+      "learning_rate": 1.4968152866242039e-05,
+      "loss": 0.0197,
+      "step": 11000
+    },
+    {
+      "epoch": 73.25,
+      "learning_rate": 1.337579617834395e-05,
+      "loss": 0.0184,
+      "step": 11500
+    },
+    {
+      "epoch": 76.43,
+      "learning_rate": 1.178343949044586e-05,
+      "loss": 0.0104,
+      "step": 12000
+    },
+    {
+      "epoch": 79.62,
+      "learning_rate": 1.0191082802547772e-05,
+      "loss": 0.0094,
+      "step": 12500
+    },
+    {
+      "epoch": 82.8,
+      "learning_rate": 8.598726114649681e-06,
+      "loss": 0.0071,
+      "step": 13000
+    },
+    {
+      "epoch": 85.99,
+      "learning_rate": 7.006369426751593e-06,
+      "loss": 0.0064,
+      "step": 13500
+    },
+    {
+      "epoch": 89.17,
+      "learning_rate": 5.414012738853504e-06,
+      "loss": 0.0077,
+      "step": 14000
+    },
+    {
+      "epoch": 92.36,
+      "learning_rate": 3.821656050955414e-06,
+      "loss": 0.0052,
+      "step": 14500
+    },
+    {
+      "epoch": 95.54,
+      "learning_rate": 2.229299363057325e-06,
+      "loss": 0.0049,
+      "step": 15000
+    },
+    {
+      "epoch": 98.73,
+      "learning_rate": 6.369426751592357e-07,
+      "loss": 0.0056,
+      "step": 15500
+    },
+    {
+      "epoch": 100.0,
+      "step": 15700,
+      "total_flos": 4.8994912512e+16,
+      "train_loss": 0.15928955945619352,
+      "train_runtime": 2364.7944,
+      "train_samples_per_second": 211.435,
+      "train_steps_per_second": 6.639
     }
   ],
   "logging_steps": 500,
-  "max_steps": 7850,
-  "num_train_epochs": 50,
+  "max_steps": 15700,
+  "num_train_epochs": 100,
   "save_steps": 500,
-  "total_flos": 2.4497456256e+16,
+  "total_flos": 4.8994912512e+16,
   "trial_name": null,
   "trial_params": null
 }