G:\Github\loftr\src\loftr\loftr.py:55: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if data['hw0_i'] == data['hw1_i']: # faster & better BN convergence
G:\Github\loftr\src\loftr\loftr_module\transformer.py:89: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
assert self.d_model == feat0.size(2), "the feature number of src and transformer must be equal"
F:\Anaconda3\envs\paddle\lib\site-packages\opt_einsum\contract.py:231: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
dim = int(sh[cnum])
F:\Anaconda3\envs\paddle\lib\site-packages\opt_einsum\parser.py:155: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
max(shape[loc] for shape, loc in zip(shapes, [x.find(c) for x in inputs]) if loc >= 0) for c in output)
F:\Anaconda3\envs\paddle\lib\site-packages\einops\einops.py:316: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
known: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] != _unknown_axis_length}
F:\Anaconda3\envs\paddle\lib\site-packages\einops\einops.py:317: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
unknown: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] == _unknown_axis_length}
G:\Github\loftr\src\loftr\loftr_module\fine_preprocess.py:35: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if data['b_ids'].shape[0] == 0:
G:\Github\loftr\src\loftr\loftr.py:81: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if feat_f0_unfold.size(0) != 0: # at least one coarse level predicted
G:\Github\loftr\src\loftr\utils\fine_matching.py:28: TracerWarning: Converting a tensor to a Python float might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
W = int(math.sqrt(WW))
G:\Github\loftr\src\loftr\utils\fine_matching.py:33: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
if M == 0:
G:\Github\loftr\src\loftr\utils\fine_matching.py:66: TracerWarning: Using len to get tensor shape might cause the trace to be incorrect. Recommended usage would be tensor.shape[0]. Passing a tensor of different shape might lead to errors or silently give incorrect results.
mkpts1_f = data['mkpts1_c'] + (coords_normed (W // 2) scale1)[:len(data['mconf'])]
The traced function didn't return any values! Side-effects are not captured in traces, so it would be a no-op.
环境： 1.paddlepaddle-gpu 2.2.2.post101 paddle2onnx 1.0.6 x2paddle 1.4.1 onnx 1.14.0 torch 2.0.0 2.Windows10 3.Python 3.10.6 4.CUDA 11.6 CUDNN 8.2
模型来源： https://github.com/zju3dv/LoFTR
下面是模型转化的时候遇到的问题：
G:\Github\loftr\src\loftr\loftr.py:55: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if data['hw0_i'] == data['hw1_i']: # faster & better BN convergence G:\Github\loftr\src\loftr\loftr_module\transformer.py:89: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! assert self.d_model == feat0.size(2), "the feature number of src and transformer must be equal" F:\Anaconda3\envs\paddle\lib\site-packages\opt_einsum\contract.py:231: TracerWarning: Converting a tensor to a Python integer might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! dim = int(sh[cnum]) F:\Anaconda3\envs\paddle\lib\site-packages\opt_einsum\parser.py:155: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! max(shape[loc] for shape, loc in zip(shapes, [x.find(c) for x in inputs]) if loc >= 0) for c in output) F:\Anaconda3\envs\paddle\lib\site-packages\einops\einops.py:316: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! known: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] != _unknown_axis_length} F:\Anaconda3\envs\paddle\lib\site-packages\einops\einops.py:317: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! unknown: Set[str] = {axis for axis in composite_axis if axis_name2known_length[axis] == _unknown_axis_length} G:\Github\loftr\src\loftr\loftr_module\fine_preprocess.py:35: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if data['b_ids'].shape[0] == 0: G:\Github\loftr\src\loftr\loftr.py:81: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if feat_f0_unfold.size(0) != 0: # at least one coarse level predicted G:\Github\loftr\src\loftr\utils\fine_matching.py:28: TracerWarning: Converting a tensor to a Python float might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! W = int(math.sqrt(WW)) G:\Github\loftr\src\loftr\utils\fine_matching.py:33: TracerWarning: Converting a tensor to a Python boolean might cause the trace to be incorrect. We can't record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs! if M == 0: G:\Github\loftr\src\loftr\utils\fine_matching.py:66: TracerWarning: Using len to get tensor shape might cause the trace to be incorrect. Recommended usage would be tensor.shape[0]. Passing a tensor of different shape might lead to errors or silently give incorrect results. mkpts1_f = data['mkpts1_c'] + (coords_normed (W // 2) scale1)[:len(data['mconf'])] The traced function didn't return any values! Side-effects are not captured in traces, so it would be a no-op.
联系方式：cswei1874@163.com