我按照NPUYOLO V3模型转换的步骤, 去转换yamnet(google开源音频分类模型)模型, 生成的inputmeta.yml如下
# !!!This file disallow TABs!!!
# "category" allowed values: "image, frequency, undefined"
# "database" allowed types: "TEXT, NPY, H5FS, SQLITE, LMDB, GENERATOR, ZIP"
# "tensor_name" only support in H5FS database
# "preproc_type" allowed types:"IMAGE_RGB, IMAGE_RGB888_PLANAR, IMAGE_RGB888_PLANAR_SEP, IMAGE_I420,
# IMAGE_NV12, IMAGE_YUV444, IMAGE_GRAY, IMAGE_BGRA, TENSOR"
input_meta:
databases:
- path: dataset.txt
type: TEXT
ports:
- lid: waveform_94
category: image
dtype: float32
sparse: false
tensor_name:
layout: nchw
shape:
- 1
fitting: scale
preprocess:
reverse_channel: true
scale: 0.0039
preproc_node_params:
add_preproc_node: false
preproc_type: TENSOR
preproc_perm:
- 0
redirect_to_output: false
显示的category是image, 但是我没有管他继续运行, 最后得到的结果是
D *********** Setup database (1) ***********
D Setup dataset "TEXT":
D Lids: ['waveform_94']
D Layouts: ['nchw']
D Shapes: [[1]]
D Data types: ['float32']
D Sparse tensors: []
D Tensor names(H5FS only): []
W:tensorflow:AutoGraph could not transform <acuitylib.core.execution.BackendFunction object at 0x7f6b5f5f42b0> and will run it as-is.
Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.
Cause: Unable to locate the source code of <function BackendFunction.__call__ at 0x7f6b605b49d0>. Note that functions defined in certain environments, like the interactive Python shell do not expose their source code. If that is the case, you should to define them in a .py source file. If you are certain the code is graph-compatible, wrap the call using @tf.autograph.do_not_convert. Original error: could not get source code
To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert
W The dimension of input0 != input1, please make sure that the version of tensorflow is greater than 1.14
I Running 1 iterations
Traceback (most recent call last):
File "pegasus.py", line 131, in <module>
File "pegasus.py", line 108, in main
File "acuitylib/app/medusa/commands.py", line 212, in execute
File "acuitylib/vsi_nn.py", line 592, in quantize
File "acuitylib/app/medusa/quantization.py", line 151, in run
File "acuitylib/app/medusa/quantization.py", line 54, in _run_quantization
File "acuitylib/app/medusa/workspace.py", line 121, in run
File "acuitylib/app/medusa/workspace.py", line 127, in _run
File "acuitylib/app/medusa/workspace.py", line 107, in _run_iteration
File "acuitylib/core/execution.py", line 567, in __call__
File "acuitylib/core/execution.py", line 506, in _dataset_feed_data
File "acuitylib/dataset/text_dataset.py", line 45, in batch
File "acuitylib/dataset/file_path_dataset.py", line 64, in _to_tensor
File "acuitylib/dataset/file_path_dataset.py", line 26, in gen_read_file_graph
File "tensorflow/python/util/dispatch.py", line 201, in wrapper
File "tensorflow/python/ops/image_ops_impl.py", line 2706, in decode_image
File "tensorflow/python/util/dispatch.py", line 201, in wrapper
File "tensorflow/python/util/deprecation.py", line 507, in new_func
File "tensorflow/python/ops/control_flow_ops.py", line 1210, in cond
File "tensorflow/python/ops/image_ops_impl.py", line 2689, in check_png
File "tensorflow/python/util/dispatch.py", line 201, in wrapper
File "tensorflow/python/util/deprecation.py", line 507, in new_func
File "tensorflow/python/ops/control_flow_ops.py", line 1210, in cond
File "tensorflow/python/ops/image_ops_impl.py", line 2676, in check_gif
File "tensorflow/python/util/dispatch.py", line 201, in wrapper
File "tensorflow/python/util/deprecation.py", line 507, in new_func
File "tensorflow/python/ops/control_flow_ops.py", line 1210, in cond
File "tensorflow/python/ops/image_ops_impl.py", line 2647, in _bmp
File "tensorflow/python/util/dispatch.py", line 201, in wrapper
File "tensorflow/python/util/tf_should_use.py", line 247, in wrapped
File "tensorflow/python/ops/control_flow_ops.py", line 154, in Assert
tensorflow.python.framework.errors_impl.InvalidArgumentError: Expected 'tf.Tensor(False, shape=(), dtype=bool)' to be true. Summarized data: b'Unable to decode bytes as JPEG, PNG, GIF, or BMP'
[4142] Failed to execute script 'pegasus' due to unhandled exception!
就算把category的image改成audio, 结果也是一样