解决Yolo2训练后不输出zip文件问题 🛠️

官方yolo2训练完不输出zip文件

2021-07-164144

Ericius

UID: 27087

1文章

0粉丝

0获取喜欢

AI 快速预览详细

本文解决了在训练Yolo2模型后无法输出zip文件的问题。具体报错信息为'FusedBatchNormV3' op既不是自定义操作也不是flex操作。通过详细的错误调试经验，提供了具体的解决方案。

在我训练完模型后，无法在out文件夹中输出zip文件，请各位大神看看，错误信息如下：

<unknown>:0: error: loc(callsite("FusedBatchNormV3@mobilenet_0.75_224_conv_pw_13_bn_cond_false_frozen_38185" at callsite("mobilenet_0.75_224/conv_pw_13_bn/cond"("/home/eric/.local/lib/python3.8/site-packages/tensorflow/python/keras/saving/hdf5_format.py":180:0) at callsite("/home/eric/.local/lib/python3.8/site-packages/tensorflow/python/keras/saving/save.py":201:0 at callsite("/home/eric/.local/lib/python3.8/site-packages/tensorflow/lite/python/lite.py":1726:0 at callsite("/home/eric/.local/lib/python3.8/site-packages/tensorflow/lite/python/lite.py":2207:0 at callsite("/mnt/d/vscode_py/maix_train/train/detector/yolo/backend/utils/fit.py":80:0 at callsite("/mnt/d/vscode_py/maix_train/train/detector/yolo/backend/utils/fit.py":56:0 at callsite("/mnt/d/vscode_py/maix_train/train/detector/yolo/frontend.py":145:0 at callsite("/mnt/d/vscode_py/maix_train/train/detector/__init__.py":200:0 at callsite("/mnt/d/vscode_py/maix_train/train/__init__.py":272:0 at "/mnt/d/vscode_py/maix_train/train/__init__.py":131:0))))))))))): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
/home/eric/.local/lib/python3.8/site-packages/tensorflow/python/keras/saving/hdf5_format.py:180:0: note: called from/home/eric/.local/lib/python3.8/site-packages/tensorflow/python/keras/saving/save.py:201:0: note: called from
/home/eric/.local/lib/python3.8/site-packages/tensorflow/lite/python/lite.py:1726:0: note: called from
/home/eric/.local/lib/python3.8/site-packages/tensorflow/lite/python/lite.py:2207:0: note: called from
/mnt/d/vscode_py/maix_train/train/detector/yolo/backend/utils/fit.py:80:0: note: called from
/mnt/d/vscode_py/maix_train/train/detector/yolo/backend/utils/fit.py:56:0: note: called from
/mnt/d/vscode_py/maix_train/train/detector/yolo/frontend.py:145:0: note: called from
/mnt/d/vscode_py/maix_train/train/detector/__init__.py:200:0: note: called from
/mnt/d/vscode_py/maix_train/train/__init__.py:272:0: note: called from
/mnt/d/vscode_py/maix_train/train/__init__.py:131:0: note: called from
复制代码

评论(1)

作者

最后还有一句
[code]:0: error: failed while converting: 'main':
Some ops are not supported by the native TFLite runtime, you can enable TF kernels fallback using TF Select. See instructions:
https://www.tensorflow.org/lite/guide/ops_select
TF Select ops: FusedBatchNormV3
Details:
tf.FusedBatchNormV3 {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}[/code]

- 没有更多了 -