eskayML/interview_classifier

README.md

@@ -4,6 +4,8 @@ license: apache-2.0
 base_model: distilbert-base-uncased
 tags:
 - generated_from_trainer
+metrics:
+- accuracy
 model-index:
 - name: interview_classifier
   results: []
@@ -15,6 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
 # interview_classifier
 
 This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.0881
+- Accuracy: 0.2593
 
 ## Model description
 
@@ -41,6 +46,22 @@ The following hyperparameters were used during training:
 - lr_scheduler_type: linear
 - num_epochs: 10
 
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss | Accuracy |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|
+| No log        | 1.0   | 54   | 2.2885          | 0.1481   |
+| No log        | 2.0   | 108  | 2.2611          | 0.1481   |
+| No log        | 3.0   | 162  | 2.2186          | 0.2593   |
+| No log        | 4.0   | 216  | 2.1877          | 0.2222   |
+| No log        | 5.0   | 270  | 2.1593          | 0.2593   |
+| No log        | 6.0   | 324  | 2.1332          | 0.2593   |
+| No log        | 7.0   | 378  | 2.1185          | 0.2963   |
+| No log        | 8.0   | 432  | 2.0965          | 0.2593   |
+| No log        | 9.0   | 486  | 2.0914          | 0.2593   |
+| 1.9418        | 10.0  | 540  | 2.0881          | 0.2593   |
+
+
 ### Framework versions
 
 - Transformers 4.46.2

config.json

@@ -38,6 +38,7 @@
   "n_heads": 12,
   "n_layers": 6,
   "pad_token_id": 0,
+  "problem_type": "single_label_classification",
   "qa_dropout": 0.1,
   "seq_classif_dropout": 0.2,
   "sinusoidal_pos_embds": false,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:21d46da7190cee32bf6b5982b0afa63c01df020db1ee3eb240e63338c06d2360
+oid sha256:aa594325c82c5854c2c695f89e782dff58b6e24bb8b4965d18940f62c2bc12e2
 size 267857176