charlie880624/STDC
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

doc: update stdc_step_by_step: some typos and minor fixes

Browse Source
This commit is contained in:
chingning.chen 2022-04-15 14:34:53 +08:00
parent b6108b81cb
commit c447bdc86d
1 changed files with 26 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
docs_kneron/stdc_step_by_step.md
View File

@ -44,7 +44,7 @@ git clone https://github.com/kneron/kneron-mmsegmentation.git
cd kneron-mmsegmentation cd kneron-mmsegmentation
``` ```
### Step 1-2-4: Install Required Python Packages for Building and Installing kneron-mmsegmentation ### Step 1-2-4: Install Required Python Libraries for Building and Installing kneron-mmsegmentation
```shell ```shell
pip install -r requirements_kneron.txt pip install -r requirements_kneron.txt
@ -83,7 +83,7 @@ ln -s $(realpath path/to/your/cityscapes) kneron-mmsegmentation/data
# Replace all "path/to/your" above with where you want to put the dataset! # Replace all "path/to/your" above with where you want to put the dataset!
``` ```
Then, we need *cityscapesScripts* to preprocess the CityScapes dataset. If you completely followed our [Step 1-2-4](#step-1-2-4-install-required-python-packages-for-building-and-installing-kneron-mmsegmentation), you should have python package *cityscapesScripts* installed (if no, execute `pip install cityscapesScripts` command). Then, we need *cityscapesScripts* to preprocess the CityScapes dataset. If you completely followed our [Step 1-2-4](#step-1-2-4-install-required-python-libraries-for-building-and-installing-kneron-mmsegmentation), you should have python library *cityscapesScripts* installed (if no, execute `pip install cityscapesScripts` command).
```shell ```shell
# Replace "path/to/your" with where you want to put the dataset! # Replace "path/to/your" with where you want to put the dataset!
@ -156,7 +156,7 @@ cd kneron-mmsegmentation
python tools/train.py configs/stdc/kn_stdc1_in1k-pre_512x1024_80k_cityscapes.py python tools/train.py configs/stdc/kn_stdc1_in1k-pre_512x1024_80k_cityscapes.py
``` ```
And kneron-mmsegmentation will generate `work_dirs/kn_stdc1_in1k-pre_512x1024_80k_cityscapes` folder and save the configuration file and all checkpoints there. kneron-mmsegmentation will generate `work_dirs/kn_stdc1_in1k-pre_512x1024_80k_cityscapes` folder and save the configuration file and all checkpoints there.
# Step 3: Test Trained Model # Step 3: Test Trained Model
`tools/test.py` is a script that generates inference results from test set with our pytorch model and evaluates the results to see if our pytorch model is well trained (if `--eval` argument is given). Note that it's always good to evluate our pytorch model before deploying it. `tools/test.py` is a script that generates inference results from test set with our pytorch model and evaluates the results to see if our pytorch model is well trained (if `--eval` argument is given). Note that it's always good to evluate our pytorch model before deploying it.
@ -241,19 +241,19 @@ Note that the ONNX results may differ from the PyTorch results due to some imple
# Step 5: Convert ONNX File to [NEF](http://doc.kneron.com/docs/#toolchain/manual/#5-nef-workflow) Model for Kneron Platform # Step 5: Convert ONNX File to [NEF](http://doc.kneron.com/docs/#toolchain/manual/#5-nef-workflow) Model for Kneron Platform
### Step 5-1: Install Kneron toolchain docker: ## Step 5-1: Install Kneron toolchain docker:
* check [document](http://doc.kneron.com/docs/#toolchain/manual/#1-installation) * Check [Kneron Toolchain Installation Document](http://doc.kneron.com/docs/#toolchain/manual/#1-installation)
### Step 5-2: Mout Kneron toolchain docker ## Step 5-2: Mount Kneron toolchain docker
* Mount a folder (e.g. '/mnt/hgfs/Competition') to toolchain docker container as `/data1`. The converted ONNX in Step 3 should be put here. All the toolchain operation should happen in this folder. * Mount a folder (e.g. '/mnt/hgfs/Competition') to toolchain docker container as `/data1`. The converted ONNX in Step 3 should be put here. All the toolchain operation should happen in this folder.
``` ```
sudo docker run --rm -it -v /mnt/hgfs/Competition:/data1 kneron/toolchain:latest sudo docker run --rm -it -v /mnt/hgfs/Competition:/data1 kneron/toolchain:latest
``` ```
### Step 5-3: Import KTC and required lib in python shell ## Step 5-3: Import KTC and the required libraries in python
* Here we demonstrate how to go through all Kneron Toolchain (KTC) flow through Python API:
```python ```python
import ktc import ktc
import numpy as np import numpy as np
@ -262,7 +262,8 @@ import onnx
from PIL import Image from PIL import Image
``` ```
### Step 5-4: Optimize the onnx model ## Step 5-4: Optimize the onnx model
```python ```python
onnx_path = '/data1/latest.onnx' onnx_path = '/data1/latest.onnx'
m = onnx.load(onnx_path) m = onnx.load(onnx_path)
@ -270,7 +271,7 @@ m = ktc.onnx_optimizer.onnx2onnx_flow(m)
onnx.save(m,'latest.opt.onnx') onnx.save(m,'latest.opt.onnx')
``` ```
### Step 5-5: Configure and load data necessary for ktc, and check if onnx is ok for toolchain ## Step 5-5: Configure and load data needed for ktc, and check if onnx is ok for toolchain
```python ```python
# npu (only) performance simulation # npu (only) performance simulation
km = ktc.ModelConfig((&)model_id_on_public_field, "0001", "720", onnx_model=m) km = ktc.ModelConfig((&)model_id_on_public_field, "0001", "720", onnx_model=m)
@ -278,13 +279,13 @@ eval_result = km.evaluate()
print("\nNpu performance evaluation result:\n" + str(eval_result)) print("\nNpu performance evaluation result:\n" + str(eval_result))
``` ```
### Step 5-6: quantize the onnx model ## Step 5-6: Quantize the onnx model
We [sampled 3 images from Cityscapes dataset](https://www.kneron.com/tw/support/education-center/?folder=OpenMMLab%20Kneron%20Edition/misc/&download=41) (3 images) as quantization data. To test our quantized model: We [sampled 3 images from Cityscapes dataset](https://www.kneron.com/tw/support/education-center/?folder=OpenMMLab%20Kneron%20Edition/misc/&download=41) (3 images) as quantization data. To test our quantized model:
1. Download the zip file 1. Download the [zip file](https://www.kneron.com/tw/support/education-center/?folder=OpenMMLab%20Kneron%20Edition/misc/&download=41)
2. Extract the zip file as a folder named `cityscapes_minitest` 2. Extract the zip file as a folder named `cityscapes_minitest`
3. Put the `cityscapes_minitest` into docker mounted folder (the path in docker container should be `/data1/cityscapes_minitest`) 3. Put the `cityscapes_minitest` into docker mounted folder (the path in docker container should be `/data1/cityscapes_minitest`)
The following script will do some preprocess(should be the same as training code) on our quantization data, and put it in a list: The following script will preprocess (should be the same as training code) our quantization data, and put it in a list:
```python ```python
import os import os
@ -303,7 +304,7 @@ for (dirpath, dirnames, filenames) in walk("/data1/cityscapes_minitest"):
img_list.append(img_data) img_list.append(img_data)
``` ```
Then perform quantization. The BIE model will be generated at `/data1/output.bie`. Then perform quantization. The generated BIE model will put generated at `/data1/output.bie`.
```python ```python
# fixed-point analysis # fixed-point analysis
@ -311,7 +312,7 @@ bie_model_path = km.analysis({"input": img_list})
print("\nFixed-point analysis done. Save bie model to '" + str(bie_model_path) + "'") print("\nFixed-point analysis done. Save bie model to '" + str(bie_model_path) + "'")
``` ```
### Step 5-7: Compile ## Step 5-7: Compile
The final step is compile the BIE model into an NEF model. The final step is compile the BIE model into an NEF model.
```python ```python
@ -343,23 +344,23 @@ Recommend you read [Kneron PLUS official document](http://doc.kneron.com/docs/#p
### Step 7-2: Download STDC example code ### Step 7-2: Download STDC example code
* Go to [Kneron education center](https://www.kneron.com/tw/support/education-center/) * Go to [Kneron education center](https://www.kneron.com/tw/support/education-center/)
* Scroll down to OpenMMLab Kneron Edition table * Scroll down to **OpenMMLab Kneron Edition** table
* Select kneron-mmsegmentation * Select **kneron-mmsegmentation**
* Select STDC * Select **STDC**
* Download stdc_plus_demo.zip * Download **stdc_plus_demo.zip**
* unzip downloaded `stdc_plus_demo` * unzip downloaded **stdc_plus_demo**
### Step 7-3: Test enviroment is ready (require [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1)) ### Step 7-3: Test enviroment is ready (require [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1))
In `stdc_plus_demo`, we provide a stdc example model and image for quick test. In `stdc_plus_demo`, we provide a STDC-Seg example model and image for quick test.
* Plug in [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1) into your computer USB port * Plug in [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1) into your computer USB port
* Go to the stdc_plus_demo folder * Go to the stdc_plus_demo folder
```bash ```bash
cd /PATH/TO/stdc_plus_demo cd /PATH/TO/stdc_plus_demo
``` ```
* Install required library * Install required python libraries
```bash ```bash
pip insall -r requirements.txt pip install -r requirements.txt
``` ```
* Run example on [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1) * Run example on [KL720 USB accelerator](https://www.kneo.ai/products/hardwares/HW2020122500000007/1)
@ -368,7 +369,7 @@ python KL720DemoGenericInferenceSTDC_BypassHwPreProc.py -nef ./example_stdc_720.
``` ```
Then you can see the inference result is saved as output_000000000641.jpg in the same folder. Then you can see the inference result is saved as output_000000000641.jpg in the same folder.
And the expected result of the command above will be something similar to the following text: The expected result of the command above will be something similar to the following text:
```plain ```plain
... ...
[Connect Device] [Connect Device]
@ -393,4 +394,4 @@ And the expected result of the command above will be something similar to the fo
Use the same script in previous step, but now we change the input NEF model path and image to yours Use the same script in previous step, but now we change the input NEF model path and image to yours
```bash ```bash
python KL720DemoGenericInferenceSTDC_BypassHwPreProc.py -img /PATH/TO/YOUR_IMAGE.bmp -nef /PATH/TO/YOUR/720_NEF_MODEL.nef python KL720DemoGenericInferenceSTDC_BypassHwPreProc.py -img /PATH/TO/YOUR_IMAGE.bmp -nef /PATH/TO/YOUR/720_NEF_MODEL.nef
``` ```