minor test

            right_pad_tokens = (new_input_ids[b] == pad_token_id).sum().item()

            # Assert that initial padding tokens have attention mask of 0

            assert torch.all(attention_mask[b, -right_pad_tokens:] == 0), "right padding tokens must have attention mask of 0"

            if not torch.all(attention_mask[b, -right_pad_tokens:] == 0):

                print("[DEBUG]: right padding tokens must have attention mask of 0")

            # Find special token indices

            sptk_b_indices = (input_ids[b] == sptk_b).nonzero().flatten()

                hole_pos.append(start_pos.item())

                hole_pos.append(end_pos.item())

            hole_pos.append(seq_len-right_pad_tokens)

            assert new_input_ids[b][seq_len-right_pad_tokens]==pad_token_id

            # assert new_input_ids[b][seq_len-right_pad_tokens]==pad_token_id

            if not torch.all(new_input_ids[b][seq_len-right_pad_tokens:] == pad_token_id):

                print("[DEBUG]: right padding tokens must be pad token")

            # shift right to fill the wholes

            holes_to_fill=1