Spaces:
Configuration error

englert commited on
Commit
1f88a07
·
1 Parent(s): a8b501d

update app.py and resnet50.py

Browse files
Files changed (2) hide show
  1. app.py +6 -4
  2. resnet50.py +1 -1
app.py CHANGED
@@ -34,8 +34,8 @@ def predict(input_file, downsample_size):
34
 
35
  zip_path = os.path.join(input_file.split('/')[-1][:-4] + ".zip")
36
 
37
- mean = np.asarray([0.3156024, 0.33569682, 0.34337464])
38
- std = np.asarray([0.16568947, 0.17827448, 0.18925823])
39
 
40
  img_vecs = []
41
  with torch.no_grad():
@@ -46,8 +46,9 @@ def predict(input_file, downsample_size):
46
  to_rgb=True)):
47
  in_img = (in_img.astype(np.float32) / 255.)
48
  in_img = (in_img - mean) / std
 
49
  in_img = np.transpose(in_img, (0, 3, 1, 2))
50
- in_img = torch.from_numpy(in_img)
51
  encoded = avg_pool(model(in_img))[0, :, 0, 0].cpu().numpy()
52
  img_vecs += [encoded]
53
 
@@ -82,7 +83,8 @@ def predict(input_file, downsample_size):
82
 
83
  demo = gr.Interface(
84
  fn=predict,
85
- inputs=[gr.inputs.Video(label="Upload Video File"), gr.inputs.Number(label="Downsample size")],
 
86
  outputs=gr.outputs.File(label="Zip"))
87
 
88
  demo.launch()
 
34
 
35
  zip_path = os.path.join(input_file.split('/')[-1][:-4] + ".zip")
36
 
37
+ mean = np.asarray([0.3156024, 0.33569682, 0.34337464], dtype=np.float32)
38
+ std = np.asarray([0.16568947, 0.17827448, 0.18925823], dtype=np.float32)
39
 
40
  img_vecs = []
41
  with torch.no_grad():
 
46
  to_rgb=True)):
47
  in_img = (in_img.astype(np.float32) / 255.)
48
  in_img = (in_img - mean) / std
49
+ in_img = np.expand_dims(in_img, 0)
50
  in_img = np.transpose(in_img, (0, 3, 1, 2))
51
+ in_img = torch.from_numpy(in_img).float()
52
  encoded = avg_pool(model(in_img))[0, :, 0, 0].cpu().numpy()
53
  img_vecs += [encoded]
54
 
 
83
 
84
  demo = gr.Interface(
85
  fn=predict,
86
+ inputs=[gr.inputs.Video(label="Upload Video File"),
87
+ gr.inputs.Number(label="Downsample size")],
88
  outputs=gr.outputs.File(label="Zip"))
89
 
90
  demo.launch()
resnet50.py CHANGED
@@ -314,7 +314,7 @@ class ResNet(nn.Module):
314
  )[1], 0)
315
  start_idx = 0
316
  for end_idx in idx_crops:
317
- _out = self.forward_backbone(torch.cat(inputs[start_idx: end_idx]).cuda(non_blocking=True))
318
  if start_idx == 0:
319
  output = _out
320
  else:
 
314
  )[1], 0)
315
  start_idx = 0
316
  for end_idx in idx_crops:
317
+ _out = self.forward_backbone(torch.cat(inputs[start_idx: end_idx])) # .cuda(non_blocking=True)
318
  if start_idx == 0:
319
  output = _out
320
  else: