-
您好,冒昧打扰。 之前拜读了您团队的Pretrained ImageProcessingTransformer, 受益匪浅。在尝试复现您的论文,但是在Dataloader和Datasets的写过程中遇到了问题。不知如何裁剪图片, 如何对数据进行预处理。论文中提到Overlap10Pix, 但是边缘部分如何处理!
若能分享解惑,不甚感激。
感谢感谢,期待回复。
-
## Environment
### Hardware Environment(`Ascend`/`GPU`/`CPU`):
`/device cpu`
### Software Environment:
- **MindSpore version (source or binary)**: `v1.2.0`
- **Python version (e.g., Python 3…
-
## Environment
### Hardware Environment(`Ascend`/`GPU`/`CPU`):
`/device cpu`
### Software Environment:
- **MindSpore version (source or binary)**: `v1.1.1`, `v1.2.0-rc`
- **Python versio…
-
## Environment
### Hardware Environment(`Ascend`/`GPU`/`CPU`):
- CPU: AMD 2950x
- GPU: not used
### Software Environment:
- **MindSpore version (source or binary)**: 1.2.0rc1
- **Python versi…
-
## Environment
### Hardware Environment(/`GPU RTX1060`/):
### Software Environment:
- **AKG version (source)**:
- **Python version 3.7.5)**:
- **OS platform and distribution (Linux Ubuntu 18.0…
wxyhv updated
3 years ago
-
想请问一下,tnn在海思端的arm有做过专门的优化吗,比如hisi3519
如果想获得海思arm端的速度极限,是不是通过 #73 中的编译选项来做?也就是:
- add_definitions( -mfpu=neon-vfpv4 -mfloat-abi=softfp )
是否还需要其他的编译选项,目前我的测试结果,在海思3519端测试mobilenetv2模型还有其他我自己训练…
-
### Version of Singularity:3.8.0-rc.1
### Expected behavior
```bash
INFO: Converting SIF file to temporary sandbox...
Tue Jun 1 10:59:58 2021
+---------------------------------------…
-
## Environment
### Hardware Environment(`Ascend`/`GPU`/`CPU`):
`/device cpu`
### Software Environment:
- **MindSpore version (source or binary)**: `v1.1.1`
- **Python version (e.g., Pyth…
-
## Environment
### Hardware Environment(`Ascend`/`GPU`/`CPU`):
GPU
8 GTX2080Ti
### Software Environment:
- **MindSpore version (source or binary)**:1.1.1 docker
- **Python version (e.g., …
-
I tried to re-implement the Deeplabv3 model on the NVIDIA GPU platform with MindSpore. However, I was surprised to find that the `ResizeBilinear` operation does not support the GPU platform. How shoul…