TVM-Quant-Graph
收藏arXiv2025-09-30 收录
下载链接：
https://discuss.tvm.apache.org/t/tf-lite-quantized-conv2d-operator-conversion/2651/30
下载链接
链接失效反馈官方服务：
资源简介：
该数据集主要关注量化技术在TVM中的表现，特别是针对conv2d策略，以及它们对推理时间和内存效率的影响。此外，数据集还包含了各种优化策略及其对计算受限和内存受限任务的影响。其任务是对量化性能进行分析。
This dataset primarily focuses on the performance of quantization techniques within TVM, particularly for the conv2d strategy, alongside the impacts of these techniques on inference latency and memory efficiency. Additionally, it covers a range of optimization strategies and their respective effects on compute-bound and memory-bound workloads. The core task of this dataset is to conduct analysis on quantization performance.
提供机构：
TVM Project
