記錄分爲
- pytorch2onnx
- onnx2tnn
- tnn結果驗證
- 移動端(安卓)使用
1、pytorch2onnx
環境:
pytorch 1.4.0
onnx 1.6.0 (轉換)
onnxruntime 1.3.0 (測試)
onnx-simplifier 0.2.9 (模型量化,不執行後續報錯了,我測試是這樣的)
轉換代碼:
import onnx
import torch
from test_net import TestModel
import numpy as np
import cv2
if 1:
torch_model = TestModel("model.pt")
torch_model.eval()
batch_size = 1 #批處理大小
input_shape = (3,384,384) #輸入數據
# set the model to inference mode
# torch_model.eval()
x = torch.randn(batch_size,*input_shape) # 生成張量
export_onnx_file = "./model.onnx" # 目的ONNX文件名
torch.onnx.export(torch_model,
x,
export_onnx_file,
export_params=True,
opset_version=11,
do_constant_folding=True, # wether to execute constant folding for optimization
input_names = ['input'], # the model's input names
output_names = ['output'], # the model's output names
dynamic_axes={'input' : {0 : 'batch_size'}, # variable lenght axes
'output' : {0 : 'batch_size'}}
)
print ('get onnx ok!')
利用 onnxruntime 測試轉換的模型:
import onnxruntime
import imageio
import time
(width, height) = (384,384)
cap = cv2.VideoCapture(0)
while 1:
ret,img = cap.read()
time_start = time.time()
if img is None:
print('no image input!')
break
if img.ndim == 2:
img = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
in_height ,in_width ,_ = img.shape
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) / 255.0
img_resized = cv2.resize(img, (width, height), interpolation=cv2.INTER_AREA)
img_resized = (
torch.from_numpy(np.transpose(img_resized, (2, 0, 1))).contiguous().float()
)
value = img_resized.unsqueeze(0)
def to_numpy(tensor):
return tensor.detach().cpu().numpy() if tensor.requires_grad else tensor.cpu().numpy()
ort_session = onnxruntime.InferenceSession("model.onnx")
ort_inputs = {ort_session.get_inputs()[0].name: (to_numpy(value)).astype(np.float32)}
#Actual: (N11onnxruntime17PrimitiveDataTypeIdEE) , expected: (N11onnxruntime17PrimitiveDataTypeIfEE)
#傳入數據類型不對
ort_outs = ort_session.run(None, ort_inputs)
result = ort_outs[0][0, 0, :, :]
result = np.array(result)
print (result.shape)
reslut_resized = cv2.resize(
result, (in_width, in_height), interpolation=cv2.INTER_AREA
)
print('cost : %.3f (s)'%(time.time() - time_start))
cv2.namedWindow('re',2)
cv2.imshow('re',reslut_resized)
if cv2.waitKey(1) ==27:
break
cap.release()
cv2.destroyAllWindows()
模型簡化操作:
python -m onnxsim model10.onnx model_sim.onnx --input-shape 1,3,384,384
2、onnx2tnn
- 下載tnn源碼,https://github.com/Tencent/TNN ;進入
~/TNN-master/tools/onnx2tnn/onnx-converter 文件夾,運行 ./build 進行編譯。
2.運行命令進行轉換
python onnx2tnn.py model/model_sim.onnx -version=algo_version -optimize=1
0.----onnx version:1.6.0
結果爲:
algo_optimize 1
onnx_net_opt_path /home/jiang/TNN-master/tools/onnx2tnn/onnx-converter/model/model_sim.opt.onnx
1.----onnx_optimizer: /home/jiang/TNN-master/tools/onnx2tnn/onnx-converter/model/model_sim.onnx
/home/jiang/TNN-master/tools/onnx2tnn/onnx-converter
----load onnx model: /home/jiang/TNN-master/tools/onnx2tnn/onnx-converter/model/model_sim.onnx
----onnxsim.simplify error: You'd better check the result with Netron
----onnxsim.simplify error: <class 'RuntimeError'>
----export optimized onnx model: /home/jiang/TNN-master/tools/onnx2tnn/onnx-converter/model/model_sim.opt.onnx
----export optimized onnx model done
2.----onnx2tnn: /home/jiang/TNN-master/tools/onnx2tnn/onnx-converter/model/model_sim.opt.onnx
get_node_attr_ai [Line 116] name :546
get_node_attr_ai [Line 116] name :585
get_node_attr_ai [Line 116] name :624
get_node_attr_ai [Line 116] name :663
get_node_attr_ai [Line 116] name :693
TNNLayerParam [Line 61] resize: coordinate_transformation_mode(pytorch_half_pixel) is not supported, result may be different.
3.----onnx2tnn status: 0
出現了錯誤----onnxsim.simplify error: You’d better check the result with Netron
----onnxsim.simplify error: <class ‘RuntimeError’>。
參數 algo_optimize=0即不進行優化就不保存轉換成功。
3、tnn結果驗證
1)、TNN編譯,不同平臺庫編譯見鏈接。以下爲安卓庫編譯:
環境要求
- 依賴庫
cmake(使用3.6及以上版本)
sudo apt-get install attr
-
NDK配置
下載ndk版本(>=15c) https://developer.android.com/ndk/downloads
配置環境變量 :
sudo gedit ~/.bashrc
#ndk
export ANDROID_NDK=/home/jiang/Android/android-ndk-r21b
export PATH=ANDROID_NDK
#tnn
export TNN_ROOT_PATH=/home/jiang/TNN-master export
PATH=TNN_ROOT_PATH
source ~/.bashrc
- 編譯
切換到~/TNN-master/scripts下,修改腳本 build_android.sh;然後再執行./build_android.sh進行編譯。
ABIA32=“armeabi-v7a with NEON”
ABIA64=“arm64-v8a”
STL=“c++_static”
SHARED_LIB=“ON” # ON表示編譯動態庫,OFF表示編譯靜態庫
ARM=“ON” # ON表示編譯帶有Arm CPU版本的庫
OPENMP=“ON” # ON表示打開OpenMP
OPENCL=“ON” # ON表示編譯帶有Arm GPU版本的庫
SHARING_MEM_WITH_OPENGL=0 # 1表示OpenGL的Texture可以與OpenCL共享
編譯完成後,在當前目錄的release目錄下生成對應的armeabi-v7a庫,arm64-v8a庫和include頭文件。
2)tnn模型驗證