Hello ,
I was trying this model on one image in DIOR dataset. I am facing this error:
`---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
in <cell line: 7>()
6
7 with torch.no_grad():
----> 8 pred_bbox = model(image, masks, word_id, word_mask)
9
10 # Convert to (x1, y1, x2, y2) format
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
---------------------------------------------------------------------------'
And thats after putting the args.size as 640 , the main problem is fixed when I place in the letterbox function 320 instead of 640 as third argument . However , although the error is removed , the output is wrong(seen by visualizing)
I used the same image and phrase you used.
Moreover, each time I rerun the model on the same input , I get different bbox.
Any ideas?
Hello , I was trying this model on one image in DIOR dataset. I am facing this error: `---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
3 frames
/content/RSVG-pytorch/models/CNN_MGVLF.py in forward(self, fv, fl) 166 167 pos = self.pos(x0).to(x0.dtype) --> 168 mask = torch.zeros([bs, x0.shape[2]]).cuda() 169 mask = mask.bool() 170 out = self.transformer(x0, mask, pos)
RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with
TORCH_USE_CUDA_DSA
to enable device-side assertions. ---------------------------------------------------------------------------' And thats after putting the args.size as 640 , the main problem is fixed when I place in the letterbox function 320 instead of 640 as third argument . However , although the error is removed , the output is wrong(seen by visualizing) I used the same image and phrase you used. Moreover, each time I rerun the model on the same input , I get different bbox. Any ideas?