Hugging Face Evaluator

how can I use max-length with task evluatro? below is my code:

task_evaluator = evaluator("text-classification")
eval_results = task_evaluator.compute(
        model_or_pipeline=model,
        data=dataset,
        tokenizer=tokenizer,
        device=args.device,
        metric=evaluate.combine([
            evaluate.load('accuracy'), 
            ConfiguredMetric(evaluate.load('f1'), average='macro'),
            ConfiguredMetric(evaluate.load('precision'), average='macro'),
            ConfiguredMetric(evaluate.load('recall'), average='macro'),
        ]),
        label_mapping=eval(map_labels)
    )```

You need to set that on the tokenizer, IIUC what you are trying to do.

1 Like