| {"xnli_ar": {"GPT-3 style_armt": {"arguments": "Namespace(config_name=None, dataset_config_name='ar', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ar', template_name='GPT-3 style_armt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ar", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "GPT-3 style_armt"}, "MNLI crowdsource_armt": {"arguments": "Namespace(config_name=None, dataset_config_name='ar', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ar', template_name='MNLI crowdsource_armt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ar", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3550200803212851}, "template_name": "MNLI crowdsource_armt"}, "can we infer_armt": {"arguments": "Namespace(config_name=None, dataset_config_name='ar', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ar', template_name='can we infer_armt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ar", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "can we infer_armt"}, "guaranteed/possible/impossible_armt": {"arguments": "Namespace(config_name=None, dataset_config_name='ar', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ar', template_name='guaranteed/possible/impossible_armt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ar", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3453815261044177}, "template_name": "guaranteed/possible/impossible_armt"}, "justified in saying_armt": {"arguments": "Namespace(config_name=None, dataset_config_name='ar', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ar', template_name='justified in saying_armt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ar", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "justified in saying_armt"}}, "xnli_es": {"GPT-3 style_esmt": {"arguments": "Namespace(config_name=None, dataset_config_name='es', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='es', template_name='GPT-3 style_esmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "es", "dataset_name": "xnli", "evaluation": {"accuracy": 0.5771084337349398}, "template_name": "GPT-3 style_esmt"}, "MNLI crowdsource_esmt": {"arguments": "Namespace(config_name=None, dataset_config_name='es', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='es', template_name='MNLI crowdsource_esmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "es", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3353413654618474}, "template_name": "MNLI crowdsource_esmt"}, "can we infer_esmt": {"arguments": "Namespace(config_name=None, dataset_config_name='es', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='es', template_name='can we infer_esmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "es", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "can we infer_esmt"}, "guaranteed/possible/impossible_esmt": {"arguments": "Namespace(config_name=None, dataset_config_name='es', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='es', template_name='guaranteed/possible/impossible_esmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "es", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3337349397590361}, "template_name": "guaranteed/possible/impossible_esmt"}, "justified in saying_esmt": {"arguments": "Namespace(config_name=None, dataset_config_name='es', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='es', template_name='justified in saying_esmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "es", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "justified in saying_esmt"}}, "xnli_fr": {"GPT-3 style_frmt": {"arguments": "Namespace(config_name=None, dataset_config_name='fr', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='fr', template_name='GPT-3 style_frmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "fr", "dataset_name": "xnli", "evaluation": {"accuracy": 0.4598393574297189}, "template_name": "GPT-3 style_frmt"}, "MNLI crowdsource_frmt": {"arguments": "Namespace(config_name=None, dataset_config_name='fr', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='fr', template_name='MNLI crowdsource_frmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "fr", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "MNLI crowdsource_frmt"}, "can we infer_frmt": {"arguments": "Namespace(config_name=None, dataset_config_name='fr', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='fr', template_name='can we infer_frmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "fr", "dataset_name": "xnli", "evaluation": {"accuracy": 0.5791164658634538}, "template_name": "can we infer_frmt"}, "guaranteed/possible/impossible_frmt": {"arguments": "Namespace(config_name=None, dataset_config_name='fr', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='fr', template_name='guaranteed/possible/impossible_frmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "fr", "dataset_name": "xnli", "evaluation": {"accuracy": 0.46947791164658637}, "template_name": "guaranteed/possible/impossible_frmt"}, "justified in saying_frmt": {"arguments": "Namespace(config_name=None, dataset_config_name='fr', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='fr', template_name='justified in saying_frmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "fr", "dataset_name": "xnli", "evaluation": {"accuracy": 0.5542168674698795}, "template_name": "justified in saying_frmt"}}, "xnli_hi": {"GPT-3 style_himt": {"arguments": "Namespace(config_name=None, dataset_config_name='hi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='hi', template_name='GPT-3 style_himt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "hi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.342570281124498}, "template_name": "GPT-3 style_himt"}, "MNLI crowdsource_himt": {"arguments": "Namespace(config_name=None, dataset_config_name='hi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='hi', template_name='MNLI crowdsource_himt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "hi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "MNLI crowdsource_himt"}, "can we infer_himt": {"arguments": "Namespace(config_name=None, dataset_config_name='hi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='hi', template_name='can we infer_himt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "hi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.41606425702811245}, "template_name": "can we infer_himt"}, "guaranteed/possible/impossible_himt": {"arguments": "Namespace(config_name=None, dataset_config_name='hi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='hi', template_name='guaranteed/possible/impossible_himt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "hi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.385140562248996}, "template_name": "guaranteed/possible/impossible_himt"}, "justified in saying_himt": {"arguments": "Namespace(config_name=None, dataset_config_name='hi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='hi', template_name='justified in saying_himt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "hi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3927710843373494}, "template_name": "justified in saying_himt"}}, "xnli_sw": {"GPT-3 style_swmt": {"arguments": "Namespace(config_name=None, dataset_config_name='sw', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='sw', template_name='GPT-3 style_swmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "sw", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3582329317269076}, "template_name": "GPT-3 style_swmt"}, "MNLI crowdsource_swmt": {"arguments": "Namespace(config_name=None, dataset_config_name='sw', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='sw', template_name='MNLI crowdsource_swmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "sw", "dataset_name": "xnli", "evaluation": {"accuracy": 0.348995983935743}, "template_name": "MNLI crowdsource_swmt"}, "can we infer_swmt": {"arguments": "Namespace(config_name=None, dataset_config_name='sw', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='sw', template_name='can we infer_swmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "sw", "dataset_name": "xnli", "evaluation": {"accuracy": 0.36265060240963853}, "template_name": "can we infer_swmt"}, "guaranteed/possible/impossible_swmt": {"arguments": "Namespace(config_name=None, dataset_config_name='sw', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='sw', template_name='guaranteed/possible/impossible_swmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "sw", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3389558232931727}, "template_name": "guaranteed/possible/impossible_swmt"}, "justified in saying_swmt": {"arguments": "Namespace(config_name=None, dataset_config_name='sw', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='sw', template_name='justified in saying_swmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "sw", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3582329317269076}, "template_name": "justified in saying_swmt"}}, "xnli_ur": {"GPT-3 style_urmt": {"arguments": "Namespace(config_name=None, dataset_config_name='ur', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ur', template_name='GPT-3 style_urmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ur", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3309236947791165}, "template_name": "GPT-3 style_urmt"}, "MNLI crowdsource_urmt": {"arguments": "Namespace(config_name=None, dataset_config_name='ur', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ur', template_name='MNLI crowdsource_urmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ur", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "MNLI crowdsource_urmt"}, "can we infer_urmt": {"arguments": "Namespace(config_name=None, dataset_config_name='ur', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ur', template_name='can we infer_urmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ur", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3409638554216867}, "template_name": "can we infer_urmt"}, "guaranteed/possible/impossible_urmt": {"arguments": "Namespace(config_name=None, dataset_config_name='ur', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ur', template_name='guaranteed/possible/impossible_urmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ur", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "guaranteed/possible/impossible_urmt"}, "justified in saying_urmt": {"arguments": "Namespace(config_name=None, dataset_config_name='ur', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='ur', template_name='justified in saying_urmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "ur", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3550200803212851}, "template_name": "justified in saying_urmt"}}, "xnli_vi": {"GPT-3 style_vimt": {"arguments": "Namespace(config_name=None, dataset_config_name='vi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='vi', template_name='GPT-3 style_vimt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "vi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "GPT-3 style_vimt"}, "MNLI crowdsource_vimt": {"arguments": "Namespace(config_name=None, dataset_config_name='vi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='vi', template_name='MNLI crowdsource_vimt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "vi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "MNLI crowdsource_vimt"}, "can we infer_vimt": {"arguments": "Namespace(config_name=None, dataset_config_name='vi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='vi', template_name='can we infer_vimt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "vi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "can we infer_vimt"}, "guaranteed/possible/impossible_vimt": {"arguments": "Namespace(config_name=None, dataset_config_name='vi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='vi', template_name='guaranteed/possible/impossible_vimt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "vi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.39196787148594375}, "template_name": "guaranteed/possible/impossible_vimt"}, "justified in saying_vimt": {"arguments": "Namespace(config_name=None, dataset_config_name='vi', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='vi', template_name='justified in saying_vimt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "vi", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3333333333333333}, "template_name": "justified in saying_vimt"}}, "xnli_zh": {"GPT-3 style_zhmt": {"arguments": "Namespace(config_name=None, dataset_config_name='zh', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='zh', template_name='GPT-3 style_zhmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "zh", "dataset_name": "xnli", "evaluation": {"accuracy": 0.5060240963855421}, "template_name": "GPT-3 style_zhmt"}, "MNLI crowdsource_zhmt": {"arguments": "Namespace(config_name=None, dataset_config_name='zh', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='zh', template_name='MNLI crowdsource_zhmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "zh", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3437751004016064}, "template_name": "MNLI crowdsource_zhmt"}, "can we infer_zhmt": {"arguments": "Namespace(config_name=None, dataset_config_name='zh', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='zh', template_name='can we infer_zhmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "zh", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3421686746987952}, "template_name": "can we infer_zhmt"}, "guaranteed/possible/impossible_zhmt": {"arguments": "Namespace(config_name=None, dataset_config_name='zh', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='zh', template_name='guaranteed/possible/impossible_zhmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "zh", "dataset_name": "xnli", "evaluation": {"accuracy": 0.42811244979919677}, "template_name": "guaranteed/possible/impossible_zhmt"}, "justified in saying_zhmt": {"arguments": "Namespace(config_name=None, dataset_config_name='zh', dataset_name='xnli', debug=False, dtype='bfloat16', max_length=2048, model_name_or_path='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498', output_dir='/gpfsscratch/rech/six/commun/commun/experiments/muennighoff/bloomckpt/176bt0/xp3capmixnewcodelonglossseq_global_step498/evaluation', pad_to_max_length=False, per_device_eval_batch_size=8, prefixlm=False, split='validation', target_max_length=256, template_config_name='zh', template_name='justified in saying_zhmt', tokenizer_name=None, use_slow_tokenizer=False)", "dataset_config_name": "zh", "dataset_name": "xnli", "evaluation": {"accuracy": 0.3377510040160643}, "template_name": "justified in saying_zhmt"}}} |