| 项目 | 版本/内容 |
|---|---|
| 设备 | Ascend 910B |
dinov3-vitb16-pretrain-lvd1689m-ascend/ # 适配代码
├── inference.py # 推理脚本
└── README.md # 本文档cd /opt/atomgit/mxy/dinov3-vitb16-pretrain-lvd1689m-ascend
python inference.py --model_path /opt/atomgit/mxy/dinov3-vitb16-pretrain-lvd1689m --precision_testcd /opt/atomgit/mxy/dinov3-vitb16-pretrain-lvd1689m-ascend
python inference.py --model_path /opt/atomgit/mxy/dinov3-vitb16-pretrain-lvd1689m --image_path /tmp/test_image.jpg --device npu:0| 参数 | 说明 | 默认值 |
|---|---|---|
| --model_path | 模型权重路径 | 必需 |
| --image_path | 待推理图像路径 | 必需(精度测试时不需要) |
| --precision_test | 运行精度测试 | False |
| --device | 运行设备 | npu:0 |
========================================================
Precision Comparison: CPU vs NPU
========================================================
Max errors: sum=9.16e-05, mean=1.19e-07, std=2.98e-08
PASS: NPU precision within thresholds
========================================================
PRECISION TEST PASSED
========================================================| 指标 | 阈值 | 实测值 | 状态 |
|---|---|---|---|
| max_error_sum | < 1e-3 | 9.16e-05 | ✅ PASS |
| max_error_mean | < 1e-5 | 1.19e-07 | ✅ PASS |
| max_error_std | < 1e-5 | 2.98e-08 | ✅ PASS |
2026-05-11 07:13:14,968 - INFO - DINOv3-ViT-B16 昇腾 NPU 推理
2026-05-11 07:13:17,549 - INFO - 模型已加载到设备: npu:0
2026-05-11 07:13:17,549 - INFO - 开始预热...
2026-05-11 07:13:17,860 - INFO - 预热完成
2026-05-11 07:13:17,860 - INFO - 开始推理...
2026-05-11 07:13:17,862 - INFO - 推理耗时: 0.0178s
2026-05-11 07:13:17,862 - INFO - 池化输出形状: torch.Size([1, 768])
2026-05-11 07:13:17,862 - INFO - 隐藏状态形状: torch.Size([1, 201, 768])
2026-05-11 07:13:17,862 - INFO - 推理成功完成!torch.npu.is_available() 返回 True，8卡均正常