Lixing He
Lixing He
It comes from image_data_augmentation of Dataset.py.
> 请问楼主解决了吗？ 其实就是中途退出触发except，在调试的时候很常见。
By the way, I have tried the solutions on: https://stackoverflow.com/questions/23284473/fatal-error-eigen-dense-no-such-file-or-directory, but seems it doesn't work.
Maybe This work few-shot training involve base and novel class, while incremental object detection doesn't. Seems the latter one is more challenging. Just my two cents, please correct me if...
For training, I believe the downsample is automatical.
@fakufaku Hello, my result shows that "only ism, max_order=-1" and "only ray tracing" have very different results, even though the room is just a simple shoebox. Do you have any...
This is the result of room.plot_rir when max_order = -1  This is the result of room.plot_rir when room.set_ray_tracing() & max_order = -1  It is clearly that they are...
Dear Xiang, Thank you for explanations！
seems you need to go to the Nvidia official repo and use ColMap to generate the transform.json, which makes your dataset aligned to the provided ones.
Same. When I downsample *4, the inference time even increase a little bit, I am digging in with some toy example, but anyone who know it please point it out.