DDP icon indicating copy to clipboard operation
DDP copied to clipboard

Results 11 DDP issues
Sort by recently updated
recently updated
newest added

Hello, Thank you very much for sharing the great work! I was trying to reproduce the segmentation result using bash tools/dist_train.sh configs/ade/ddp_swin_s_2x8_512x512_160k_ade20k.py 1 Unfortunately, I cannot reproduce a similar result...

Thank you for your excellent work! I want to train the model using my own dataset, how does this modify the configuration file? How should I train.

I encountered a confusing problem when I tried to train the depth model. When I executed `bash tools/dist_train.sh configs/ddp_kitti/ddp_swinb_22k_w7_kitti_bs2x8_scale01.py 1` with samples_per_gpu=2, the model took the following GPU memory. ![image](https://github.com/JiYuanFeng/DDP/assets/80506854/4efe2435-a363-42d1-886b-25b7d3257bc2)...

Hello, I am getting following error: nvrtc: error: invalid value for --gpu-architecture (-arch) While trying to train the depth estimation model using the following command: bash tools/dist_train.sh configs/custom/custom_config.py 1 I...

When I run the segmentation code on other public datasets, the loss becomes nan after training for more than 20 epochs. What could be the reason? Thank you for your...

Hello, thanks for your work! when train the depth model, it shows error:'ModuleNotFoundError: No module named 'depth.models.depther.regulardepth''. I do not find the 'regulardepth' module in your code. Could you solve...

KITTI's depth gt is a map with sparse ground truth values. Do you use it directly as x0 of the diffusion model during training?

I first fork bevfusion and replace configs and mmdet3d with ddp's configs and mmdet3d. When training ddp, I encountered this error: - RuntimeError: "ms_deform_attn_forward_cuda" not implemented for 'Half' I try...

Hi, When I try to run inference on a demo image using the image_demo.py file, I find that the key: 'pad_shape' is missing. How do I resolve this issue? Thanks

Thanks for your great work! In segmention/mmseg/models/deformable_head_with_time.py, the input augments are "query","key","value","time" , etc in self.encoder function, line 119. But, in mmcv package, https://github.com/open-mmlab/mmcv/blob/v1.6.2/mmcv/ops/multi_scale_deform_attn.py, The "MultiScaleDeformableAttention" does not accept the...