Back to Megatron Lm

tensor_parallel package

docs/api-guide/core/tensor_parallel.md

23.06798 B
Original Source
<!--- Copyright (c) 2022-2026, NVIDIA CORPORATION. All rights reserved. NVIDIA CORPORATION and its licensors retain all intellectual property and proprietary rights in and to this software, related documentation and any modifications thereto. Any use, reproduction, disclosure or distribution of this software and related documentation without an express license agreement from NVIDIA CORPORATION is strictly prohibited. -->

tensor_parallel package

This package contains an implementation for tensor parallelism in transformer models (see Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism and Reducing Activation Recomputation in Large Transformer Models for details).