__MHGL__

I get an exciting idea, i try to save freeze_model to disk, then use torch.jit.load get freeze_model before convert coreml, this problem not happened! like this: ```python3 import torch import...

[pytorch docs](https://pytorch.org/docs/stable/fx.html#limitations-of-symbolic-tracing)

# coreml转换问题 ## numpy版本问题 - numpy == 1.21.0 - coremltools == 0.4.1 - error ``` File "/home/liyang/HomePod/deepvac/deepvac/cast/coreml.py", line 60, in process coreml_model = coremltools.convert(model=model, inputs=[input], File "/home/liyang/.local/lib/python3.8/site-packages/coremltools/converters/_converters_entry.py", line 175, in...

> 单纯的5次下采样并不会引发这个错误，还需要在最后的feature上上采样+cat操作。 > 这个问题很奇怪，有时可以正常运行，会出现`[TensorRT] WARNING: Max value of this profile is not valid`，就只是warning；有的时候就是直接报错了。暂时没有能力深究原因，等下官方吧。

rewrite Focus ```python3 from torch.nn import functional as F class Focus(torch.nn.Module): def __init__(self): super(Focus, self).__init__() ... def forward(self, x): # x(b,c,w,h) -> y(b,4c,w/2,h/2) # part-1 gain = torch.tensor([[1, 0], [0,...

```python3 from coremltools.converters.mil import register_torch_op from coremltools.converters.mil.frontend.torch.ops import _get_inputs, _get_scales_from_output_size from coremltools.converters.mil.mil import Builder as mb @register_torch_op(override=True) def upsample_bilinear2d(context, node): inputs = _get_inputs(context, node) _input = inputs[0] output_size = inputs[1]...

@bkykent it works, thank u

感谢您的回复。我尝试过编译插件so库（通过pytorch cuda extension 或者 CMakeLists中类似libdetector），这方面确实经验比较少，您可以给点建议吗？

同问，511和271有什么依据吗？

``` I RKNN: [13:27:50.940] RKNN Runtime Information, librknnrt version: 1.6.0 (9a7b5d24c@2023-12-13T17:31:11) I RKNN: [13:27:50.940] RKNN Driver Information, version: 0.8.8 I RKNN: [13:27:50.940] RKNN Model Information, version: 6, toolkit version: 1.5.2+b642f30c(compiler...