Qia98
Qia98
请问单阶段网络（如config里面的YOLOv3）的蒸馏策略是怎么样的？teacher和student如何选取？以及损失函数如何处理？
I want to train 512*512 resolution on the voc dataset. Can someone give me some help, such as croping or filtering the vox dataset to 512 resolution
When I train the 512 model, I noticed that the vis picture seems to have been cropped
@SystemErrorWang How to use distributed data parallel to accelerate the training ？I put the model and datasets to DDP. but it seams to be not working. the GPU usage was...
> > performance killer, I have fix for the fixed shape situation: > > > > * [nonezero](https://github.com/IDEA-Research/GroundingDINO/blob/main/groundingdino/models/GroundingDINO/bertwarper.py#L240) > > * [arange](https://github.com/IDEA-Research/GroundingDINO/blob/main/groundingdino/models/GroundingDINO/bertwarper.py#L256) > > HIi @oylz ! I met this...
> > > performance killer, I have fix for the fixed shape situation: > > > > > > * [nonezero](https://github.com/IDEA-Research/GroundingDINO/blob/main/groundingdino/models/GroundingDINO/bertwarper.py#L240) > > > * [arange](https://github.com/IDEA-Research/GroundingDINO/blob/main/groundingdino/models/GroundingDINO/bertwarper.py#L256) > > > >...
And I use this command try sanitise it: ``` ./polygraphy surgeon sanitize --fold-constants ./input.onnx -o output.onnx ``` result is ``` [I] RUNNING | Command: /miniconda3/envs/tensorrt/bin/polygraphy surgeon sanitize gd_token_no_dy_sigmoid_512_fold.onnx --fold-constants -o...
I met the same problem, did you solve it? It looks like a data set format problem
> I use yt-dlp，it works fine. I've tried yt-dlp, but it didn't work. Could you share the voxceleb1 dataset (id 10000-20000) with me? My emali is [email protected]. I've already obtained...