metadata

language:
  - en
thumbnail: >-
  https://avatars3.githubusercontent.com/u/32437151?s=460&u=4ec59abc8d21d5feea3dab323d23a5860e6996a4&v=4
tags:
  - text-classification
  - ag_news
  - pytorch
license: mit
datasets:
  - ag_news
metrics:
  - accuracy
model-index:
  - name: nateraw/bert-base-uncased-ag-news
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: ag_news
          type: ag_news
          config: default
          split: test
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9414473684210526
            verified: true
          - name: Precision Macro
            type: precision
            value: 0.9416084922682436
            verified: true
          - name: Precision Micro
            type: precision
            value: 0.9414473684210526
            verified: true
          - name: Precision Weighted
            type: precision
            value: 0.9416084922682435
            verified: true
          - name: Recall Macro
            type: recall
            value: 0.9414473684210527
            verified: true
          - name: Recall Micro
            type: recall
            value: 0.9414473684210526
            verified: true
          - name: Recall Weighted
            type: recall
            value: 0.9414473684210526
            verified: true
          - name: F1 Macro
            type: f1
            value: 0.9414706154045142
            verified: true
          - name: F1 Micro
            type: f1
            value: 0.9414473684210526
            verified: true
          - name: F1 Weighted
            type: f1
            value: 0.9414706154045143
            verified: true
          - name: loss
            type: loss
            value: 0.17173650860786438
            verified: true

bert-base-uncased-ag-news

Model description

bert-base-uncased finetuned on the AG News dataset using PyTorch Lightning. Sequence length 128, learning rate 2e-5, batch size 32, 4 T4 GPUs, 4 epochs. The code can be found here

Limitations and bias

Not the best model...

Training data

Data came from HuggingFace's datasets package. The data can be viewed on nlp viewer.

Training procedure

...

Eval results

...