Upload trained sentiment classifier

README.md

@@ -4,6 +4,11 @@ license: mit
 base_model: xlm-roberta-base
 tags:
 - generated_from_trainer
+metrics:
+- accuracy
+- f1
+- precision
+- recall
 model-index:
 - name: bert-sentiment-classifier
   results: []
@@ -15,6 +20,15 @@ should probably proofread and complete it, then remove this comment. -->
 # bert-sentiment-classifier
 
 This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1918
+- Accuracy: 0.935
+- F1: 0.9350
+- Precision: 0.9349
+- Recall: 0.9351
+- F1 Negative: 0.9362
+- F1 Neutral: 0.9337
+- F1 Positive: 0.0
 
 ## Model description
 

all_results.json

@@ -8,12 +8,12 @@
     "eval_loss": 0.19182628393173218,
     "eval_precision": 0.9349253582055963,
     "eval_recall": 0.9350576454944666,
-    "eval_runtime": 3.8282,
-    "eval_samples_per_second": 522.443,
-    "eval_steps_per_second": 2.09,
+    "eval_runtime": 3.8007,
+    "eval_samples_per_second": 526.217,
+    "eval_steps_per_second": 2.105,
     "total_flos": 1303200758006784.0,
     "train_loss": 0.4567002405094195,
-    "train_runtime": 90.5072,
-    "train_samples_per_second": 110.488,
-    "train_steps_per_second": 0.873
+    "train_runtime": 90.6217,
+    "train_samples_per_second": 110.349,
+    "train_steps_per_second": 0.872
 }

eval_results.json

@@ -8,7 +8,7 @@
     "eval_loss": 0.19182628393173218,
     "eval_precision": 0.9349253582055963,
     "eval_recall": 0.9350576454944666,
-    "eval_runtime": 3.8282,
-    "eval_samples_per_second": 522.443,
-    "eval_steps_per_second": 2.09
+    "eval_runtime": 3.8007,
+    "eval_samples_per_second": 526.217,
+    "eval_steps_per_second": 2.105
 }

train_results.json

@@ -2,7 +2,7 @@
     "epoch": 1.0,
     "total_flos": 1303200758006784.0,
     "train_loss": 0.4567002405094195,
-    "train_runtime": 90.5072,
-    "train_samples_per_second": 110.488,
-    "train_steps_per_second": 0.873
+    "train_runtime": 90.6217,
+    "train_samples_per_second": 110.349,
+    "train_steps_per_second": 0.872
 }