ram02's picture
Update metadata with huggingface_hub
545d9a5 verified
metadata
datasets:
  - alinet/balanced_qg
model-index:
  - name: alinet/bart-base-balanced-resolved-qg
    results:
      - task:
          type: text2text-generation
          name: Question Generation
        dataset:
          name: MRQA
          type: mrqa
        metrics:
          - type: bertscore
            value: 0.6550437803614988
            name: BERTScore F1
          - type: bertscore
            value: 0.6511161012190818
            name: BERTScore Precision
          - type: bertscore
            value: 0.6625115818906895
            name: BERTScore Recall
      - task:
          type: text2text-generation
          name: Question Generation
        dataset:
          name: Spoken-SQuAD
          type: alinet/spoken_squad
        metrics:
          - type: bertscore
            value: 0.5983651754615461
            name: BERTScore F1
          - type: bertscore
            value: 0.5884801388565024
            name: BERTScore Precision
          - type: bertscore
            value: 0.6120697321749161
            name: BERTScore Recall

A question generation model trained on alinet/balanced_qg dataset (resolved subset).

Example usage:

from transformers import BartConfig, BartForConditionalGeneration, BartTokenizer

model_name = "alinet/bart-base-balanced-resolved-qg"

tokenizer = BartTokenizer.from_pretrained(model_name)
model = BartForConditionalGeneration.from_pretrained(model_name) 

def run_model(input_string, **generator_args):
  input_ids = tokenizer.encode(input_string, return_tensors="pt")
  res = model.generate(input_ids, **generator_args)
  output = tokenizer.batch_decode(res, skip_special_tokens=True)
  print(output)

run_model("Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.", max_length=32, num_beams=4)
# ['What is the term for a reading comprehension dataset consisting of questions posed by crowdworkers?']