gpt2_tokenizer

--- dataset_info: features: - name: input_ids sequence: int32 - name: attention_mask sequence: int8 splits: - name: train num_bytes: 16291456129 num_examples: 80462898 down