onnx转rknn失败

[1515292252]

文件夹中best.pt是源文件，v10-320-200-1122.onnx是best.pt转换成的，onnx_to_rknn.py是我自己写的转换脚本，log_feedback_to_the_rknn_toolkit_dev_team.log是运行onnx_to_rknn.py后生成的日志，报错如下：
W rknn-toolkit version: 1.7.5
--> Loading model
W base_optimize check ...
W "output0": cosine = 1.0, euclidean = 0.0
W correct_ops check ...
W "output0": cosine = 1.0, euclidean = 0.0
W fuse_ops check ...
W "output0": cosine = 1.0, euclidean = 0.0
E Calc node GatherElements : /model.23/GatherElements output shape fail
E Catch exception when loading onnx model: ./optimize.onnx!
E Traceback (most recent call last):
E File "rknn/base/RKNNlib/onnx_ir/onnx_numpy_backend/ops/gather_elements.py", line 19, in rknn.base.RKNNlib.onnx_ir.onnx_numpy_backend.ops.gather_elements.GatherElements
E File "rknn/base/RKNNlib/onnx_ir/onnx_numpy_backend/ops/gather_elements.py", line 12, in rknn.base.RKNNlib.onnx_ir.onnx_numpy_backend.ops.gather_elements.GatherElements.gather_elements
E File "<array_function internals>", line 5, in choose
E File "/home/tsh/anaconda3/envs/py38_rknn1.7.5/lib/python3.8/site-packages/numpy/core/fromnumeric.py", line 430, in choose
E return _wrapfunc(a, 'choose', choices, out=out, mode=mode)
E File "/home/tsh/anaconda3/envs/py38_rknn1.7.5/lib/python3.8/site-packages/numpy/core/fromnumeric.py", line 58, in _wrapfunc
E return bound(*args, **kwds)
E ValueError: Need at least 0 and at most 32 array objects.
E During handling of the above exception, another exception occurred:
E Traceback (most recent call last):
E File "rknn/base/RKNNlib/app/importer/import_onnx.py", line 120, in rknn.base.RKNNlib.app.importer.import_onnx.Importonnx.run
E File "rknn/base/RKNNlib/converter/convert_onnx.py", line 123, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx.init
E File "rknn/base/RKNNlib/converter/convert_onnx.py", line 1187, in rknn.base.RKNNlib.converter.convert_onnx.convert_onnx._shape_inference
E File "rknn/base/RKNNlib/onnx_ir/onnx_numpy_backend/shape_inference.py", line 70, in rknn.base.RKNNlib.onnx_ir.onnx_numpy_backend.shape_inference.infer_shape
E File "rknn/api/rknn_log.py", line 323, in rknn.api.rknn_log.RKNNLog.e
E ValueError: Calc node GatherElements : /model.23/GatherElements output shape fail
E Please feedback the detailed log file <log_feedback_to_the_rknn_toolkit_dev_team.log> to the RKNN Toolkit development team.
E You can also check github issues: https://github.com/rockchip-linux/rknn-toolkit/issues
load model failed
Exception ignored in: <function RKNN.del at 0x7f3ddf350820>
Traceback (most recent call last):
File "/home/tsh/anaconda3/envs/py38_rknn1.7.5/lib/python3.8/site-packages/rknn/api/rknn.py", line 781, in del
AttributeError: 'NoneType' object has no attribute 'FileHandler'
问GPT说是输出节点形状不对，要修改，但是也没说应该改成什么样子的呀。
这个在3588上是能跑的所以源文件应该不会有问题，但现在想部署在1126上，如何解决呢？
log_feedback_to_the_rknn_toolkit_dev_team.log

[frankyzc]

问题解决了吗

[1515292252]

问题解决了吗
还没有 你也遇到这个问题了吗

[frankyzc]

对，加个微信？ 发自我的iPhone

…

------------------ 原始邮件 ------------------ 发件人: 1515292252 ***@***.***&gt; 发送时间: 2024年12月16日 15:11 收件人: rockchip-linux/rknn-toolkit ***@***.***&gt; 抄送: frankyzc ***@***.***&gt;, Comment ***@***.***&gt; 主题: Re: [rockchip-linux/rknn-toolkit] onnx转rknn失败 (Issue #491) 问题解决了吗 还没有 你也遇到这个问题了吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***&gt;

[1515292252]

可以 tsh980612

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人: "rockchip-linux/rknn-toolkit" ***@***.***&gt;; 发送时间:&nbsp;2024年12月18日(星期三) 凌晨4:50 ***@***.***&gt;; ***@***.******@***.***&gt;; 主题:&nbsp;Re: [rockchip-linux/rknn-toolkit] onnx转rknn失败 (Issue #491) 对，加个微信？ 发自我的iPhone

------------------ 原始邮件 ------------------ 发件人: 1515292252 ***@***.***&amp;gt; 发送时间: 2024年12月16日 15:11 收件人: rockchip-linux/rknn-toolkit ***@***.***&amp;gt; 抄送: frankyzc ***@***.***&amp;gt;, Comment ***@***.***&amp;gt; 主题: Re: [rockchip-linux/rknn-toolkit] onnx转rknn失败 (Issue #491) 问题解决了吗 还没有 你也遇到这个问题了吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***&amp;gt; — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***&gt;