diff options
Diffstat (limited to 'docs/sections/troubleshooting.md')
-rw-r--r-- | docs/sections/troubleshooting.md | 31 |
1 files changed, 15 insertions, 16 deletions
diff --git a/docs/sections/troubleshooting.md b/docs/sections/troubleshooting.md index a4f60fb..8ab81eb 100644 --- a/docs/sections/troubleshooting.md +++ b/docs/sections/troubleshooting.md @@ -1,27 +1,26 @@ # Troubleshooting -- [Inference results are incorrect for my custom files](#inference-results-are-incorrect-for-my-custom-files) -- [The application does not work with my custom model](#the-application-does-not-work-with-my-custom-model) +- [Troubleshooting](#troubleshooting) + - [Inference results are incorrect for my custom files](#inference-results-are-incorrect-for-my-custom-files) + - [The application does not work with my custom model](#the-application-does-not-work-with-my-custom-model) ## Inference results are incorrect for my custom files -Ensure that the files you are using match the requirements of the model -you are using and that cmake parameters are set accordingly. More -information on these cmake parameters is detailed in their separate -sections. Note that preprocessing of the files could also affect the -inference result, such as the rescaling and padding operations done for -image classification. +Ensure that the files you are using match the requirements of the model you are using. Ensure that cmake parameters are +set accordingly. More information on these cmake parameters is detailed in their separate sections. + +> **Note:** Preprocessing of the files can also affect the inference result, such as the rescaling and padding +> operations performed for image classification. ## The application does not work with my custom model -Ensure that your model is in a fully quantized `.tflite` file format, -either uint8 or int8, and has successfully been run through the Vela -compiler. +Ensure that your model is in a fully quantized `.tflite` file format, either `uint8` or `int8`, and that it has +successfully been run through the Vela compiler. -Check that cmake parameters match your new models input requirements. +Also, please check that the cmake parameters used match the input requirements of your new model. -> **Note:** Vela tool is not available within this software project. -It is a python tool available from <https://pypi.org/project/ethos-u-vela/>. -The source code is hosted on <https://git.mlplatform.org/ml/ethos-u/ethos-u-vela.git/>. +> **Note:** The Vela tool is not available within this software project. It is a separate Python tool that is available +> from: <https://pypi.org/project/ethos-u-vela/>. The source code is hosted on +> <https://git.mlplatform.org/ml/ethos-u/ethos-u-vela.git/>. -Next section of the documentation: [Appendix](appendix.md). +The next section of the documentation is: [Appendix](appendix.md). |