I have trained a model based on Tensorflow. This model is supposed to work on the mobile phone but I have got a problem when converting froze graph (pb) to deep learning container(dlc). I have to set the input size to be constant. This cause that model can't work with any input size. I am trying to find a way that resizes input shapes of a DLC model without initializing model with "snpe-tensorflow-to-dlc --input_dims 1,512,512,3" because this way is consuming. Actually, I want to resize input shapes in dlc model. can anybody help me?
Is there any way to resize input shapes in SNPE (dlc)?
780 Views Asked by ALiBahri At
1
There are 1 best solutions below
Related Questions in TENSORFLOW
- A deterministic GPU implementation of fused batch-norm backprop, when training is disabled, is not currently available
- Keras similarity calculation. Enumerating distance between two tensors, which indicates as lists
- Does tensorflow have a way of calculating input importance for simple neural networks
- How to predict input parameters from target parameter in a machine learning model?
- Windows 10 TensorFlow cannot detect Nvidia GPU
- unable to use ignore_class in SparseCategoricalCrossentropy
- Why is this code not working? I've tried everything and everything seems to be fine, but no
- Why convert jpeg into tfrecords?
- ValueError: The shape of the target variable and the shape of the target value in `variable.assign(value)` must match
- The kernel appears to have died. It will restart automatically. whenever i try to run the plt.imshow() and plt.show() function in jupyter notebook
- Pneumonia detection, using transfer learning
- Cannot install tensorflow ver 2.3.0 (distribution not found)
- AttributeError: module 'keras._tf_keras.keras.layers' has no attribute 'experimental'
- Error while loading .keras model: Layer node index out of bounds
- prediction model with python tensorflow and keras, gives error when predicting
Related Questions in DYNAMIC
- VBA dynamic feed multiple files into current one but error of "Run-rime error 7 out of memory" occurs
- SSRS use a dynamic SQL query with parameter
- Go to the Next section in Google Forms after an option is selected using App Script
- Add and remove dynamic component Angular
- Server Side Rendering of Dynamic URL using NUXT 3
- html to PDF with new page detection
- How to absolutely position pin icons to different locations when you zoom in an image using the react-zoom-pan-pinch npm package
- Loading dynamic content for offline downloaded website
- Unable to find chart for react.js
- i want to use a dynamic expression in PIVOT values
- How do I dynamically load a CSS file in a TMS WEB Core Website using Delphi?
- Flutter DropdownButton Dynamic Default Value Error
- How to Pass-in a Collection name and Document Key to an AQL query to update the document
- Adding dynamic choices to ChoiceType form field in Symfony 6
- Pass class type that subclass or implements class; then access static methods & create instances of that type; is it possible in Java, and how?
Related Questions in RESIZE
- Change back to default font size in Android Studio
- Change the maximum screen size icon (and variable) by dragging the TitleBar
- Java resized JPEG seems to be invalid format for Facebook?
- I can't handle resize and transform with jquery
- Excel VBA Macro Resizing the Plot Area
- How to disable dragging while resizing a card using the @dnd-kit/core library in React?
- How to resize maximized child window programmatically
- CSS - making an element short on its own
- JavaFX dynamically resizes window based on the amount of VBox elements
- How can I get the width and height of a GUI window in AHK?
- Image won't resize when live but ok in Visual studio code
- QOpenGLWidget with Qpainter: OpenGL drawing only visible after first paintGL call
- ResizeObserver callback is not being fired when rotated device two times
- Addition of one 4-bit and one 3-bit inputs in VHDL
- How to resize circle div dynamically?
Related Questions in SNPE
- SNPE Model conversion: Syntax for model input with 0 dimensions
- SNPE inference outputs misaligned images
- How to modify Sample Code 1.51 to run DLC with AIP mode like snpe-net-run?
- Snapdragon neural processing- SNPE
- pb to dlc conversion with snpe fails using a re trained Tensorflow Object Detection model
- Unable to load the frozen model (.pb) in GraphDef in tensorflow version 2.x
- Is there any way to resize input shapes in SNPE (dlc)?
- SNPE - How can I predict on a model with two (or more) inputs?
- Why SNPE SDK is very Slow?
- Does Qualcomm SNPE work on a linux board?
- How to transform classify_image_graph_def.pb from "TF for poets" tutorial into inception_v3_2016_08_28_frozen.pb from Qualcomm SNPE tutorial
- Problem Converting Caffe model to dlc using SNPE
- Snapdragon Neural Processing Engine (SNPE) and Tiny Yolo on Android App
- Implementing mobilenetssd with SNPE
- ElementWise resolver must implement broadcast method - Qualcomm SNPE
Trending Questions
- UIImageView Frame Doesn't Reflect Constraints
- Is it possible to use adb commands to click on a view by finding its ID?
- How to create a new web character symbol recognizable by html/javascript?
- Why isn't my CSS3 animation smooth in Google Chrome (but very smooth on other browsers)?
- Heap Gives Page Fault
- Connect ffmpeg to Visual Studio 2008
- Both Object- and ValueAnimator jumps when Duration is set above API LvL 24
- How to avoid default initialization of objects in std::vector?
- second argument of the command line arguments in a format other than char** argv or char* argv[]
- How to improve efficiency of algorithm which generates next lexicographic permutation?
- Navigating to the another actvity app getting crash in android
- How to read the particular message format in android and store in sqlite database?
- Resetting inventory status after order is cancelled
- Efficiently compute powers of X in SSE/AVX
- Insert into an external database using ajax and php : POST 500 (Internal Server Error)
Popular Questions
- How do I undo the most recent local commits in Git?
- How can I remove a specific item from an array in JavaScript?
- How do I delete a Git branch locally and remotely?
- Find all files containing a specific text (string) on Linux?
- How do I revert a Git repository to a previous commit?
- How do I create an HTML button that acts like a link?
- How do I check out a remote Git branch?
- How do I force "git pull" to overwrite local files?
- How do I list all files of a directory?
- How to check whether a string contains a substring in JavaScript?
- How do I redirect to another webpage?
- How can I iterate over rows in a Pandas DataFrame?
- How do I convert a String to an int in Java?
- Does Python have a string 'contains' substring method?
- How do I check if a string contains a specific word?
Usually deployment solutions work with fixed input shapes because they assume some widely acknowledged usage model - resize all picture of the same certain size and do inference. And due to this usage model, developers of deployment solutions do not prioritize model loading time while they usually prioritize the inference time. The same happens in SNPE, in OpenVINO, in TFLite, etc.
To illustrate the times, here is some results from Snapdragon 820. To load Inception v3 to CPU takes 715ms, to load model to DSP takes 3 seconds. Inference on CPU takes 1 sec, inference on DSP takes 100ms. You see that loading time on DSP is bigger than on CPU, but inference time is much much better.
At the same time, usually it is allowed to change a shape before loading of the model assuming that all input pictures will have different size (but again, same for all pictures) than shapes for which model was trained. For SNPE it is SNPEBuilder::setInputDimensions If model allow to do reshape and if no bugs in SNPE implementation, the model can be reshaped and loaded.
Not sure if your usage model fits to the vision described in the first paragraph. At the same time, to have a benefit from different input size you need to develop special topology that unlikely be supported by SNPE. If you take just regular SSD and reshape it to different size and measure accuracy on validation set, the most likely you get the best result on shpaes where model was trained.