Unverified 提交 23d0456b authored 作者: Glenn Jocher's avatar Glenn Jocher 提交者: GitHub

`torch.jit.trace()` fix (#9363)

* Update common.py Signed-off-by: 's avatarGlenn Jocher <glenn.jocher@ultralytics.com> * Update ci-testing.yml Signed-off-by: 's avatarGlenn Jocher <glenn.jocher@ultralytics.com> Signed-off-by: 's avatarGlenn Jocher <glenn.jocher@ultralytics.com>
上级 cafdd189
......@@ -119,9 +119,12 @@ jobs:
python export.py --weights $m.pt --img 64 --include torchscript # export
python - <<EOF
import torch
im = torch.zeros([1, 3, 64, 64])
for path in '$m', '$b':
model = torch.hub.load('.', 'custom', path=path, source='local')
print(model('data/images/bus.jpg'))
model(im) # warmup, build grids for trace
torch.jit.trace(model, [im])
EOF
- name: Test classification
shell: bash # for Windows compatibility
......
......@@ -600,6 +600,7 @@ class AutoShape(nn.Module):
if self.pt:
m = self.model.model.model[-1] if self.dmb else self.model.model[-1] # Detect()
m.inplace = False # Detect.inplace=False for safe multithread inference
m.export = True # do not output loss values
def _apply(self, fn):
# Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论