README.md

metadata

language: en
tags:
  - summarization
model-index:
  - name: yuvraj/summarizer-cnndm
    results:
      - task:
          type: summarization
          name: Summarization
        dataset:
          name: sepidmnorozy/Urdu_sentiment
          type: sepidmnorozy/Urdu_sentiment
          config: sepidmnorozy--Urdu_sentiment
          split: train
        metrics:
          - name: ROUGE-1
            type: rouge
            value: 0
            verified: true
          - name: ROUGE-2
            type: rouge
            value: 0
            verified: true
          - name: ROUGE-L
            type: rouge
            value: 0
            verified: true
          - name: ROUGE-LSUM
            type: rouge
            value: 0
            verified: true
          - name: loss
            type: loss
            value: 10.730116844177246
            verified: true
          - name: gen_len
            type: gen_len
            value: 19.9912
            verified: true

​

Summarization

​

Model description

​ BartForConditionalGeneration model fine tuned for summarization on 10000 samples from the cnn-dailymail dataset ​

How to use

​ PyTorch model available ​

from transformers import AutoTokenizer, AutoModelWithLMHead, pipeline
​
tokenizer = AutoTokenizer.from_pretrained("yuvraj/summarizer-cnndm") 
AutoModelWithLMHead.from_pretrained("yuvraj/summarizer-cnndm")
​
summarizer = pipeline('summarization', model=model, tokenizer=tokenizer)
summarizer("<Text to be summarized>")
​
## Limitations and bias
Trained on a small dataset