docker/intel_code/llama13b/Megatron-DeepSpeed/tests/conftest.py

# Copyright (C) 2024 Habana Labs, Ltd. an Intel Company.
# Copyright (c) 2023, NVIDIA CORPORATION. All rights reserved.

import pytest

from deepspeed.accelerator import get_accelerator

from megatron.core import parallel_state
from megatron.core.tensor_parallel.random import model_parallel_cuda_manual_seed

from megatron.core.transformer.transformer_config import TransformerConfig

# import torch

# torch.distributed.init_process_group(
#     backend=get_accelerator().communication_backend_name(),
#     world_size=1, rank=1)

# # initialize model parallel for tests
# parallel_state.set_tensor_model_parallel_world_size(1)
# parallel_state.set_tensor_model_parallel_rank(0)
# parallel_state._set_global_memory_buffer()
# parallel_state.set_pipeline_model_parallel_rank(0)
# parallel_state.set_pipeline_model_parallel_world_size(1)
# parallel_state.initialize_model_parallel()

# model_parallel_cuda_manual_seed(123)


@pytest.fixture
def transformer_config():
    print(f"transformer_config")
    return TransformerConfig(num_layers=2, hidden_size=12, num_attention_heads=4, use_cpu_initialization=True)