Quantization On Imagenet
评估指标
Top-1 Accuracy (%)
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | Top-1 Accuracy (%) |
|---|---|
| training-multi-bit-quantized-and-binarized | 71.5 |
| hmq-hardware-friendly-mixed-precision | 76 |
| hmq-hardware-friendly-mixed-precision | 75.45 |
| fq-vit-fully-quantized-vision-transformer | 71.61 |
| fq-vit-fully-quantized-vision-transformer | 82.71 |
| fq-vit-fully-quantized-vision-transformer | 83.31 |
| fq-vit-fully-quantized-vision-transformer | 81.20 |
| hptq-hardware-friendly-post-training | 73.356 |
| r-2-range-regularization-for-model | 68.45 |
| hptq-hardware-friendly-post-training | 71.46 |
| learned-step-size-quantization | 76.7 |
| hptq-hardware-friendly-post-training | 74.216 |
| fq-vit-fully-quantized-vision-transformer | 80.51 |
| multi-prize-lottery-ticket-hypothesis-finding-1 | 74.03 |
| learned-step-size-quantization | 77.878 |
| fq-vit-fully-quantized-vision-transformer | 85.03 |
| r-2-range-regularization-for-model | 69.79 |
| lsq-improving-low-bit-quantization-through | 73.8 |
| learned-step-size-quantization | 77.34 |
| lsq-improving-low-bit-quantization-through | 71.7 |
| hptq-hardware-friendly-post-training | 77.092 |
| hptq-hardware-friendly-post-training | 78.972 |
| hmq-hardware-friendly-mixed-precision | 70.9 |
| hmq-hardware-friendly-mixed-precision | 76.4 |
| fq-vit-fully-quantized-vision-transformer | 82.97 |
| r-2-range-regularization-for-model | 69.64 |
| fq-vit-fully-quantized-vision-transformer | 79.17 |