evaluation_l1/anli/dev_r1/can_we_infer/results.json

{
  "dataset_name": "anli",
  "dataset_config_name": "dev_r1",
  "template_name": "can we infer",
  "evaluation": {
    "accuracy": 0.29
  },
  "arguments": "Namespace(config_name=None, dataset_config_name='dev_r1', dataset_name='anli', debug=False, dtype='float16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/experiments/muennighoff/bloomckpt/350mt0/bloomz-560m', output_dir='/gpfsscratch/rech/six/commun/experiments/muennighoff/bloomckpt/350mt0/bloomz-560m/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='dev_r1', target_max_length=256, template_config_name=None, template_name='can we infer', tokenizer_name=None, use_slow_tokenizer=False)"
}