更新文档

GAP8/ai_examples/classification_demo/README.md

@@ -1,67 +1,64 @@
-# Sample MNIST Build using GAPFlow
+# 构建数据集
 
-This project includes a sample based on a simple model graph defined in Keras.
+- 数据集组织方式：
 
-It goes from training right through to working code on GAP8 or the same code running on the PC for
-debugging purposes.
-
-* It first trains the network using keras
-* It then exports the network to H5 format
-* It then converts the H5 file to a TFLITE file using TensorFlow's TFLITE converter
-* It then generates an nntool state file by running an nntool script with commands to adjust the tensor and activation order, fuse certain operations together and automatically quantify the graph
-* It then uses this state file to generate an AutoTiler graph model
-* It then compiles and runs the model to produce GAP8 code
-* Finally it builds and runs the produced code on GAP8
-
-While this example uses Makefiles the same steps could be achieved with any build system.
+```
+\data
+   \train
+       \class1
+          \xxx.jpg
+          \xxx.jpg
+       \class2
+       ...
+       \classn
+   \test
+       \xxx.jpg
+       \xxx.jpg
+       ...
+       \xxx.jpg
+```
 
-The process can be run to quantize the model in 16 or 8 bits weights and activations.
+- data_dir = "训练集根目录"
 
-## Build and run without Docker
-*This example uses the SDK Version `3.8.1`*
+- 构建方式
 
-To build and run on GAP8:
-```
-make all run
-```
-To build and run on GVSOC
 ```
-make all run platform=gvsoc
+python dataloader.py
 ```
-The input image is specified in the Makefile and loaded with the functions defined in ${GAP_SDK_HOME}/libs/gap_lib/img_io/ImgIO.c
 
-To clean the generated model and code but not the trained network type
-```
-make clean
-```
-To clean the trained keras save file type
-```
-make clean_train
-```
-To build and run the network compiled on the pc
-```
-make -f emul.mk all
-```
-This will produce an executable, model_emul, that can be used to evaluate files
+保存生成的train.npz、val.npz文件
+
+
+# 环境要求
+
+- GAP_SDK: release-v3.8.1  (docker or ubuntu)
+- Keras: 2.3.1
+- tensorflow: 1.15.2
+- numpy: 1.16.2
+
+
+# 运行实例
+
+- 在GAP8上运行
+
 ```
-e.g. ./model_emul images/5558_6.pgm 
+make clean all run'
 ```
-This mode allows the application to be run with PC tools like valgrind which is very interesting for debugging.
-The cluster only has one core in this mode.
 
-The build defaults to 8 bit quantization. 16 bit quantization can be selected by preceeding the build lines above with QUANT_BITS=16.
+- 在gvsoc虚拟平台运行
+
 ```
-e.g. QUANT_BITS=16 make -f emul.mk all
+make clean all run platform=gvsoc
 ```
 
-## Build and run within the docker
-*This example uses the SDK Version `3.7`.*
+- 固化到芯片
 
-Navigate to top-level folder of the example
-```
-cd <path/to/ai-deck/repository>/GAP8/ai_examples/gapflow_model_extern/
 ```
-Build and execute in Docker similar as the other examples
+make clean all image flash
 ```
-docker run --rm -it -v $PWD:/module/data/ --device /dev/ttyUSB0 --privileged -P gapsdk:3.7 /bin/bash -c 'export GAPY_OPENOCD_CABLE=interface/ftdi/olimex-arm-usb-tiny-h.cfg; source /gap_sdk/configs/ai_deck.sh; cd /module/data/;  make clean clean_train clean_model clean_images all run'
+
+# 清理编译产物
+
 ```
+make clean clean_train clean_images clean_model
+```