将使用应用程序模块下载的模型（ ?weights ）转换为 tflite

Question

我正在尝试将使用 tf.keras 中的应用程序模块下载的 mobilenet 模型转换为 tensorflow lite 格式。我使用的 TensorFlow 版本是 1.31。我不知道模型实际上是仅存储权重还是权重+架构+优化器状态。当我尝试转换命令时：

from tensorflow import lite

lite.TFLiteConverter.from_keras_model_file( '/path/to/mobilenet_1_0_224_tf.h5' )

导致了这个错误：

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/lite/python/lite.py", line 370, in from_keras_model_file
    keras_model = _keras.models.load_model(model_file)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/keras/engine/saving.py", line 232, in load_model
    raise ValueError('No model found in config file.')
ValueError: No model found in config file.

据此，我假设模型只是权重。因此，我尝试使用应用程序模块加载模型，并尝试使用 model.save() 保存模型。但这导致了以下错误。

Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 300, in __init__
    fetch, allow_tensor=True, allow_operation=True))
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 3478, in as_graph_element
    return self._as_graph_element_locked(obj, allow_tensor, allow_operation)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 3557, in _as_graph_element_locked
    raise ValueError("Tensor %s is not an element of this graph." % obj)
ValueError: Tensor Tensor("conv1/kernel/Read/ReadVariableOp:0", shape=(3, 3, 3, 32), dtype=float32) is not an element of this graph.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/keras/engine/network.py", line 1334, in save
    save_model(self, filepath, overwrite, include_optimizer)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/keras/engine/saving.py", line 111, in save_model
    save_weights_to_hdf5_group(model_weights_group, model_layers)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/keras/engine/saving.py", line 742, in save_weights_to_hdf5_group
    weight_values = K.batch_get_value(symbolic_weights)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/keras/backend.py", line 2819, in batch_get_value
    return get_session().run(tensors)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1137, in _run
    self._graph, fetches, feed_dict_tensor, feed_handles=feed_handles)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 471, in __init__
    self._fetch_mapper = _FetchMapper.for_fetch(fetches)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 261, in for_fetch
    return _ListFetchMapper(fetch)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 370, in __init__
    self._mappers = [_FetchMapper.for_fetch(fetch) for fetch in fetches]
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 370, in <listcomp>
    self._mappers = [_FetchMapper.for_fetch(fetch) for fetch in fetches]
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 271, in for_fetch
    return _ElementFetchMapper(fetches, contraction_fn)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 307, in __init__
    'Tensor. (%s)' % (fetch, str(e)))
ValueError: Fetch argument <tf.Variable 'conv1/kernel:0' shape=(3, 3, 3, 32) dtype=float32> cannot be interpreted as a Tensor. (Tensor Tensor("conv1/kernel/Read/ReadVariableOp:0", shape=(3, 3, 3, 32), dtype=float32) is not an element of this graph.)

有谁知道这里真正的问题是什么？ TIA

Answer 1

你是如何保存你的模型的，也许你只保存了权重而不是模型，并且你试图调用不存在的加载模型。

如果这不是问题，请尝试清除会话。

from keras.backend import clear_session
clear_session()

我是这样转换模型的

converter = tf.lite.TFLiteConverter.from_keras_model_file('model name')
tflite_model = converter.convert()
open("converted/model.tflite", "wb").write(tflite_model)