CODEWITHSUNDEEP
androidtensorflowmachine-learningcomputer-visionartificial-intelligence
Ajinkya
Ajinkya

Reputation: 1887

Unexpected failure when preparing tensor allocations

I am using tflite for semantic segmentation. I have a model trained to segment objects from background, this model is trained on deeplab.

I have converted this model(frozen inference graph) into tflite format using the below code:

tflite_convert \
  --output_file=test.lite \
  --graph_def_file=frozen_inference_graph.pb \
  --input_arrays=ImageTensor \
  --output_arrays=SemanticPredictions \
  --input_shapes=1,600,450,3 \
  --inference_input_type=QUANTIZED_UINT8 \
  --inference_type=FLOAT \
  --mean_values=128 \
  --std_dev_values=128

The model loads on android, but when I try to run inference it gives me this error:

Caused by: java.lang.IllegalStateException: Internal error: Unexpected failure when preparing tensor allocations: third_party/tensorflow/lite/kernels/unpack.cc:54 NumDimensions(input)

1 was not true.Node number 4 (UNPACK) failed to prepare.

How do I resove this error?

Upvotes: 0

Views: 2470

Answers (1)

Ajinkya
Ajinkya

Reputation: 1887

It worked for the below command:

bazel-bin/tensorflow/lite/toco/toco \
  --input_file=deeplabv3_mnv2_pascal_tain.pb \
  --output_file=test.tflite \
  --inference_input_type=QUANTIZED_UINT8  \
  --inference_type=FLOAT \
  --input_arrays=ImageTensor  \
  --output_arrays=SemanticPredictions  \
  --input_shapes=1,513,513,3 \
  --mean_values=128 \
  --std_dev_values=128

It worked when I installed tensorflow from source. In order to install tensorflow from source I used (link)

Upvotes: 0

Related Questions