Khadas Docs

Amazing Khadas, always amazes you!

User Tools

Site Tools

Home / products / sbc / vim4 / npu / npu-sdk
products:sbc:vim4:npu:npu-sdk

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
Last revision Both sides next revision
products:sbc:vim4:npu:npu-sdk [2023/08/30 03:02]
louis 		products:sbc:vim4:npu:npu-sdk [2024/06/06 21:54]
nick
Line 1: Line 1:
 +~~tag>VIM4 NPU Docker ~~
  
-====== NPU Model Convert ======+====== VIM4 NPU Model Convert ====== 
 + 
 +{{indexmenu_n>2}}
  
 <WRAP important > <WRAP important >
-Only **New VIM4** supports NPU, you can check the version of your VIM4 here: [[products:sbc:vim4:configurations:identify-version|]]+Only **New VIM4** supports NPU, you can [[../configurations/identify-version|check VIM4 the version]]
 </WRAP> </WRAP>
  
Line 10: Line 13:
 </WRAP> </WRAP>
  
-===== Build Virtual Environment =====+===== Build Docker Environment =====
  
 Follow Docker official docs to install Docker: [[https://docs.docker.com/engine/install/ubuntu/|Install Docker Engine on Ubuntu]]. Follow Docker official docs to install Docker: [[https://docs.docker.com/engine/install/ubuntu/|Install Docker Engine on Ubuntu]].
  
-Get Docker:+Follow the script below to get Docker image:
  
 ```shell ```shell
-$ cd vim4_npu_sdk +docker pull numbqq/npu-vim4
-$ docker pull yanwyb/npu:v1 +
-$ docker run -it --name vim4-npu1 -v $(pwd):/home/khadas/npu \ +
- -v /etc/localtime:/etc/localtime:ro \ +
- -v /etc/timezone:/etc/timezone:ro \ +
- yanwyb/npu:v1+
 ``` ```
  
Line 31: Line 29:
 $ cd vim4_npu_sdk $ cd vim4_npu_sdk
 $ ls $ ls
-bin  demo  docs  README.md+adla-toolkit-binary  adla-toolkit-binary-1.2.0.9  convert-in-docker.sh  Dockerfile  docs  README.md
 ``` ```
  
-  * ''docs'' - SDK documentations +  * ''adla-toolkit-binary/docs'' - SDK documentations 
-  * ''bin'' - SDK tools required for model conversion +  * ''adla-toolkit-binary/bin'' - SDK tools required for model conversion 
-  * ''demo'' - Conversion examples+  * ''adla-toolkit-binary/demo'' - Conversion examples
  
 ===== Convert Model ===== ===== Convert Model =====
Line 43: Line 41:
  
 ```shell ```shell
-$ khadas@2655b6cbbc01:~/npu$ cd demo/ +./convert-in-docker.sh
-$ khadas@2655b6cbbc01:~/npu/demo$ bash convert_adla.sh+
 ``` ```
  
Line 50: Line 47:
  
 ```shell ```shell
-$ khadas@2655b6cbbc01:~/npu/demo$ ls+~/vim4_npu_sdk$ ls adla-toolkit-binary/demo/
 caffe_output     darknet_output  dataset.txt   libstdc++_so  mxnet_output  paddle_output   quantized_tflite_output  tflite_output caffe_output     darknet_output  dataset.txt   libstdc++_so  mxnet_output  paddle_output   quantized_tflite_output  tflite_output
 convert_adla.sh  data            keras_output  model_source  onnx_output   pytorch_output  tensorflow_output convert_adla.sh  data            keras_output  model_source  onnx_output   pytorch_output  tensorflow_output
Line 56: Line 53:
  
 Converted model is ''xxxx.adla''. Converted model is ''xxxx.adla''.
 +
 +<WRAP tip >
 +If you want to convert your own model, just modify script ''adla-toolkit-binary/demo/convert_adla.sh'' and  then execute ''./convert-in-docker.sh'' to convert your model.
 +</WRAP>
 +
  
 ==== Important parameters ==== ==== Important parameters ====
Line 63: Line 65:
   * ''inputs/input-shapes'' - ''inputs''  Input node names. ''input-shapes''  The sizes of input nodes. Need to be set when model-type is pytorch/onnx/mxnet/tensorflow/ tflite/paddle.   * ''inputs/input-shapes'' - ''inputs''  Input node names. ''input-shapes''  The sizes of input nodes. Need to be set when model-type is pytorch/onnx/mxnet/tensorflow/ tflite/paddle.
   * ''outputs'' - Output node names. When model-type is tensorflow, outputs needs to be set.   * ''outputs'' - Output node names. When model-type is tensorflow, outputs needs to be set.
-  * ''dtypes'' - Input type, set the type information corresponding to the input (optional).Default: float32. +  * ''dtypes'' - Input type, set the type information corresponding to the input (optional).Default ''float32''. 
-  * ''quantize-dtype'' - Quantization type. Currently, it supports int8, int16 and uint8 quantization types.+  * ''quantize-dtype'' - Quantization type. Currently, it supports ''int8'', ''int16'' and ''uint8'' quantification types.
   * ''source-file'' - ''dataset.txt''. The txt file contains paths to the images that need to be quantized. It supports images and npy files.   * ''source-file'' - ''dataset.txt''. The txt file contains paths to the images that need to be quantized. It supports images and npy files.
   * ''channel-mean-value'' - The pre-processing parameters are set according to the pre processing methods used during model training. It includes four values, m1, m2, m3, and scale. The first three are mean-value parameters and the last one is the scale parameter. For the input data with three channels (data1, data2, data3), the pre-processing steps are   * ''channel-mean-value'' - The pre-processing parameters are set according to the pre processing methods used during model training. It includes four values, m1, m2, m3, and scale. The first three are mean-value parameters and the last one is the scale parameter. For the input data with three channels (data1, data2, data3), the pre-processing steps are
Line 72: Line 74:
   * ''batch-size'' - Set the batch-size for the adla file after conversion. Currently, the default value is 1.   * ''batch-size'' - Set the batch-size for the adla file after conversion. Currently, the default value is 1.
   * ''iterations'' - Optional parameter, if ''dataset.txt'' provides data for multi-groups, and needs to use all data provided for quantization, set ''iterations'' and make sure iterations*batch size = Number of data in the multi-groups.   * ''iterations'' - Optional parameter, if ''dataset.txt'' provides data for multi-groups, and needs to use all data provided for quantization, set ''iterations'' and make sure iterations*batch size = Number of data in the multi-groups.
-  * ''outdir'' - Directory for the generated. The default value is the current directory ''./''.+  * ''outdir'' - Directory for the generated. The default value is the current directory.
   * ''target-platform'' - Specify the target platform for the adla file, it should be ''PRODUCT_PID0XA003'' for VIM4.   * ''target-platform'' - Specify the target platform for the adla file, it should be ''PRODUCT_PID0XA003'' for VIM4.
  
Last modified: 2024/06/13 05:30 by louis

Page Tools

Khadas Site

Khadas Community

Khadas GitHub

Khadas Downloads