End of training

README.md

@@ -5,9 +5,36 @@ tags:
 - generated_from_trainer
 datasets:
 - conll2003
+metrics:
+- precision
+- recall
+- f1
+- accuracy
 model-index:
 - name: test-ner-run
-  results: []
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: conll2003
+      type: conll2003
+      config: conll2003
+      split: validation
+      args: conll2003
+    metrics:
+    - name: Precision
+      type: precision
+      value: 0.946991165194199
+    - name: Recall
+      type: recall
+      value: 0.9560753954897341
+    - name: F1
+      type: f1
+      value: 0.9515115986935768
+    - name: Accuracy
+      type: accuracy
+      value: 0.9903625248237997
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,6 +43,12 @@ should probably proofread and complete it, then remove this comment. -->
 # test-ner-run
 
 This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the conll2003 dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0414
+- Precision: 0.9470
+- Recall: 0.9561
+- F1: 0.9515
+- Accuracy: 0.9904
 
 ## Model description
 

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 3.0,
-    "eval_accuracy": 0.9816967969394644,
-    "eval_f1": 0.6512477175897748,
-    "eval_loss": 0.08961711823940277,
-    "eval_precision": 0.6832694763729247,
-    "eval_recall": 0.622093023255814,
-    "eval_runtime": 3.8121,
-    "eval_samples": 90,
-    "eval_samples_per_second": 23.609,
-    "eval_steps_per_second": 1.574,
-    "total_flos": 660065444222976.0,
-    "train_loss": 0.0696317084930228,
-    "train_runtime": 242.0841,
-    "train_samples": 842,
-    "train_samples_per_second": 10.434,
-    "train_steps_per_second": 0.657
+    "eval_accuracy": 0.9903625248237997,
+    "eval_f1": 0.9515115986935768,
+    "eval_loss": 0.041357845067977905,
+    "eval_precision": 0.946991165194199,
+    "eval_recall": 0.9560753954897341,
+    "eval_runtime": 11.0076,
+    "eval_samples": 3250,
+    "eval_samples_per_second": 295.251,
+    "eval_steps_per_second": 18.533,
+    "total_flos": 1022109083576478.0,
+    "train_loss": 0.04506039325269498,
+    "train_runtime": 475.8923,
+    "train_samples": 14041,
+    "train_samples_per_second": 88.514,
+    "train_steps_per_second": 5.535
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 3.0,
-    "eval_accuracy": 0.9816967969394644,
-    "eval_f1": 0.6512477175897748,
-    "eval_loss": 0.08961711823940277,
-    "eval_precision": 0.6832694763729247,
-    "eval_recall": 0.622093023255814,
-    "eval_runtime": 3.8121,
-    "eval_samples": 90,
-    "eval_samples_per_second": 23.609,
-    "eval_steps_per_second": 1.574
+    "eval_accuracy": 0.9903625248237997,
+    "eval_f1": 0.9515115986935768,
+    "eval_loss": 0.041357845067977905,
+    "eval_precision": 0.946991165194199,
+    "eval_recall": 0.9560753954897341,
+    "eval_runtime": 11.0076,
+    "eval_samples": 3250,
+    "eval_samples_per_second": 295.251,
+    "eval_steps_per_second": 18.533
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.0,
-    "total_flos": 660065444222976.0,
-    "train_loss": 0.0696317084930228,
-    "train_runtime": 242.0841,
-    "train_samples": 842,
-    "train_samples_per_second": 10.434,
-    "train_steps_per_second": 0.657
+    "total_flos": 1022109083576478.0,
+    "train_loss": 0.04506039325269498,
+    "train_runtime": 475.8923,
+    "train_samples": 14041,
+    "train_samples_per_second": 88.514,
+    "train_steps_per_second": 5.535
 }

trainer_state.json

@@ -3,26 +3,56 @@
   "best_model_checkpoint": null,
   "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 159,
+  "global_step": 2634,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.050873196659074e-05,
+      "loss": 0.1325,
+      "step": 500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.1017463933181475e-05,
+      "loss": 0.047,
+      "step": 1000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.152619589977221e-05,
+      "loss": 0.0251,
+      "step": 1500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.2034927866362947e-05,
+      "loss": 0.0193,
+      "step": 2000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 2.5436598329536827e-06,
+      "loss": 0.0106,
+      "step": 2500
+    },
     {
       "epoch": 3.0,
-      "step": 159,
-      "total_flos": 660065444222976.0,
-      "train_loss": 0.0696317084930228,
-      "train_runtime": 242.0841,
-      "train_samples_per_second": 10.434,
-      "train_steps_per_second": 0.657
+      "step": 2634,
+      "total_flos": 1022109083576478.0,
+      "train_loss": 0.04506039325269498,
+      "train_runtime": 475.8923,
+      "train_samples_per_second": 88.514,
+      "train_steps_per_second": 5.535
     }
   ],
   "logging_steps": 500,
-  "max_steps": 159,
+  "max_steps": 2634,
   "num_train_epochs": 3,
   "save_steps": 500,
-  "total_flos": 660065444222976.0,
+  "total_flos": 1022109083576478.0,
   "trial_name": null,
   "trial_params": null
 }