コード例 #1
0
    def create_and_check_openai_gpt_model(self, config, input_ids, head_mask, token_type_ids, *args):
        model = OpenAIGPTModel(config=config)
        model.to(torch_device)
        model.eval()

        result = model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
        result = model(input_ids, token_type_ids=token_type_ids)
        result = model(input_ids)

        self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
コード例 #2
0
    def __init__(self, config):
        super().__init__(config)

        config.num_labels = 1
        self.transformer = OpenAIGPTModel(config)
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        self.multiple_choice_head = SequenceSummary(config)
        self.persona_head = SequenceSummary(config)

        self.init_weights()
コード例 #3
0
        def create_and_check_openai_gpt_model(self, config, input_ids, head_mask, token_type_ids, *args):
            model = OpenAIGPTModel(config=config)
            model.to(torch_device)
            model.eval()

            model(input_ids, token_type_ids=token_type_ids, head_mask=head_mask)
            model(input_ids, token_type_ids=token_type_ids)
            (sequence_output,) = model(input_ids)

            result = {"sequence_output": sequence_output}
            self.parent.assertListEqual(
                list(result["sequence_output"].size()), [self.batch_size, self.seq_length, self.hidden_size],
            )
def convert_openai_checkpoint_to_pytorch(openai_checkpoint_folder_path, openai_config_file, pytorch_dump_folder_path):
    # Construct model
    if openai_config_file == "":
        config = OpenAIGPTConfig()
    else:
        config = OpenAIGPTConfig.from_json_file(openai_config_file)
    model = OpenAIGPTModel(config)

    # Load weights from numpy
    load_tf_weights_in_openai_gpt(model, config, openai_checkpoint_folder_path)

    # Save pytorch-model
    pytorch_weights_dump_path = pytorch_dump_folder_path + '/' + WEIGHTS_NAME
    pytorch_config_dump_path = pytorch_dump_folder_path + '/' + CONFIG_NAME
    print("Save PyTorch model to {}".format(pytorch_weights_dump_path))
    torch.save(model.state_dict(), pytorch_weights_dump_path)
    print("Save configuration file to {}".format(pytorch_config_dump_path))
    with open(pytorch_config_dump_path, "w", encoding="utf-8") as f:
        f.write(config.to_json_string())