GPUMODE/backendbench_tests
收藏官方服务：
资源简介：
TorchBench数据集是BackendBench的子集，用于模拟真实世界的用例。它提供了来自TIMM、Hugging Face Transformers和TorchBench的155个模型迹中的操作符和输入。这个数据集旨在帮助开发者理解和优化PyTorch操作的性能，包含详细的字段信息，如操作符名称、参数、运行时等。
The TorchBench dataset is a subset of BackendBench designed to mimic real-world use cases. It provides operators and inputs from 155 model traces found in TIMM, Hugging Face Transformers, and TorchBench. This dataset is intended to help developers understand and optimize the performance of PyTorch operations, containing detailed information such as operator names, arguments, runtime, etc.
提供机构：
GPUMODE
