pts, pts_img = pts.view(68, 2) * 4, pts_img.view(68, 2)

` pts, pts_img, scores = self.get_preds_fromhm(out, center.numpy(), scale)
pts, pts_img = torch.from_numpy(pts), torch.from_numpy(pts_img)
pts, pts_img = pts.view(68, 2) * 4, pts_img.view(68, 2)
scores = scores.squeeze(0)`

我测试这里的pts.view(68, 2)本身返回的就是256，这里不需要乘4吧？

Look at the get_preds_fromhm function in the model to understand its output scaling. Check if pts are already in the correct coordinate space (e.g., 256x256) and whether the multiplication by 4 is necessary. Run the existing tests or create a small script to verify the landmark coordinates before and after the scaling operation.

Written by the indexing model from the issue text.