search

Xilinx / Vitis-AI-Tutorials

MIT License
395 stars 146 forks source link

14-caffe-ssd-pascal-Quantizing error #54

Open 2018boyan opened 2 years ago

2018boyan commented 2 years ago

When I run ./quantize_and_compile.sh in docker, I find this:

I0215 02:19:57.546900 259 layer_factory.hpp:77] Creating layer data I0215 02:19:57.546921 259 net.cpp:94] Creating Layer data I0215 02:19:57.546927 259 net.cpp:409] data -> data I0215 02:19:57.546942 259 net.cpp:409] data -> label I0215 02:19:57.547078 259 image_data_layer.cpp:41] Opening file ../calibration.txt I0215 02:19:57.547338 259 image_data_layer.cpp:51] Shuffling data I0215 02:19:57.547390 259 image_data_layer.cpp:56] A total of 1000 images. E0215 02:19:57.547423 259 io.cpp:145] Could not open or find file /data2/datasets/VOCdevkit/VOC2007/JPEGImages/000186.jpg F0215 02:19:57.547428 259 image_data_layer.cpp:70] Check failed: cv_img.data Could not load 000186.jpg Check failure stack trace: Compiling network: vgg16_ssd [INFO] Namespace(batchsize=1, inputs_shape=None, layout='NCHW', model_files=['quantize/deploy.caffemodel'], model_type='caffe', named_inputs_shape=None, out_filename='/tmp/vgg16_ssd_org.xmodel', proto='quantize/deploy.prototxt') [ERROR] Not found the file or directory: /workspace/SSD/VAI/VGG16-SSD/quantize/deploy.caffemodel

The 000186.jpg is in /SSD/data/VOCdevkit/VOC2007/JPEGImages/000186.jpg seems that not in /data2/datasets/VOCdevkit/VOC2007/JPEGImages/000186.jpg. It seems the wrong address.

Another question is that I want to use kv260 arch.json, but arch=/opt/vitis_ai/compiler/arch/DPUCZDX8G/ZCU102/arch.json 2>&1 | tee ${output_dir}/compile.txt in quantize_and_compile.sh could not find anything, even not has /opt/vitis_ai in my ubuntu 20.04.

Could you please help me? Thank you in advance!

bhargavin1872008 commented 2 years ago

when running the requirements.txt of keras-yolov3-modelset -i 'm getting error for coremltools.it is showing like "couldn't find a version that satisfies the requirement tensorflow<=1.14 and tensorflow >=1.5(from tfcoremltools -r requirements.txt).(from version :2.2.0,2.2..1, 2.2.2, ...2.7.0rc0,2.7.0.rc1............) like this .can someone help me regarding this. Also ,i have a doubt .can we use ubuntu 20.04 ,cuda 11.7 ,cudnn 8.4.0 for this project. or have to use ubuntu 18.04,cuda 10.0 only which only works.please help me regarding this,i have less time in my hand.

2018boyan commented 2 years ago

when running the requirements.txt of keras-yolov3-modelset -i 'm getting error for coremltools.it is showing like "couldn't find a version that satisfies the requirement tensorflow<=1.14 and tensorflow >=1.5(from tfcoremltools -r requirements.txt).(from version :2.2.0,2.2..1, 2.2.2, ...2.7.0rc0,2.7.0.rc1............) like this .can someone help me regarding this. Also ,i have a doubt .can we use ubuntu 20.04 ,cuda 11.7 ,cudnn 8.4.0 for this project. or have to use ubuntu 18.04,cuda 10.0 only which only works.please help me regarding this,i have less time in my hand.

Maybe you should change a Vitis-ai version? https://github.com/Xilinx/Vitis-AI-Tutorials/blob/1.4/Design_Tutorials/07-yolov4-tutorial/readme.md image

bhargavin1872008 commented 2 years ago

yes,im using vitis-ai 1.4 version only.but using ubuntu 20.04 and cuda 11.1 which is already installed.i dont know why im getting error for coremltools for requirements.txt of keras-yolov3-set. please spare ur valuable time.and address what is the thing wrong here. moreover i would like to seek ur help regarding the tensorflow version we are using in the python files proved by yolov4-tutorial .please help me as i am unable to differentiate tensorflow 1.x and tensorflow2.x thankyou

On Sat, Aug 20, 2022 at 10:20 AM zhouqiang2018 @.***> wrote:

when running the requirements.txt of keras-yolov3-modelset -i 'm getting error for coremltools.it is showing like "couldn't find a version that satisfies the requirement tensorflow<=1.14 and tensorflow >=1.5(from tfcoremltools -r requirements.txt).(from version :2.2.0,2.2..1, 2.2.2, ...2.7.0rc0,2.7.0.rc1............) like this .can someone help me regarding this. Also ,i have a doubt .can we use ubuntu 20.04 ,cuda 11.7 ,cudnn 8.4.0 for this project. or have to use ubuntu 18.04,cuda 10.0 only which only works.please help me regarding this,i have less time in my hand.

Maybe you should change a Vitis-ai version?

https://github.com/Xilinx/Vitis-AI-Tutorials/blob/1.4/Design_Tutorials/07-yolov4-tutorial/readme.md [image: image] https://user-images.githubusercontent.com/42407841/185729423-893d514d-fc5c-425c-9109-a9d7533499ac.png

— Reply to this email directly, view it on GitHub https://github.com/Xilinx/Vitis-AI-Tutorials/issues/54#issuecomment-1221232286, or unsubscribe https://github.com/notifications/unsubscribe-auth/A2CH5GJGSX6QNBQIRNSRUS3V2BP2RANCNFSM5ONF3L2A . You are receiving this because you commented.Message ID: @.***>

2018boyan commented 2 years ago

yes,im using vitis-ai 1.4 version only.but using ubuntu 20.04 and cuda 11.1 which is already installed.i dont know why im getting error for coremltools for requirements.txt of keras-yolov3-set. please spare ur valuable time.and address what is the thing wrong here. moreover i would like to seek ur help regarding the tensorflow version we are using in the python files proved by yolov4-tutorial .please help me as i am unable to differentiate tensorflow 1.x and tensorflow2.x thankyou … On Sat, Aug 20, 2022 at 10:20 AM zhouqiang2018 @.> wrote: when running the requirements.txt of keras-yolov3-modelset -i 'm getting error for coremltools.it is showing like "couldn't find a version that satisfies the requirement tensorflow<=1.14 and tensorflow >=1.5(from tfcoremltools -r requirements.txt).(from version :2.2.0,2.2..1, 2.2.2, ...2.7.0rc0,2.7.0.rc1............) like this .can someone help me regarding this. Also ,i have a doubt .can we use ubuntu 20.04 ,cuda 11.7 ,cudnn 8.4.0 for this project. or have to use ubuntu 18.04,cuda 10.0 only which only works.please help me regarding this,i have less time in my hand. Maybe you should change a Vitis-ai version? https://github.com/Xilinx/Vitis-AI-Tutorials/blob/1.4/Design_Tutorials/07-yolov4-tutorial/readme.md [image: image] https://user-images.githubusercontent.com/42407841/185729423-893d514d-fc5c-425c-9109-a9d7533499ac.png — Reply to this email directly, view it on GitHub <#54 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/A2CH5GJGSX6QNBQIRNSRUS3V2BP2RANCNFSM5ONF3L2A . You are receiving this because you commented.Message ID: @.> Maybe not active? https://github.com/Xilinx/Vitis-AI-Tutorials/blob/1.4/Design_Tutorials/02-MNIST_classification_tf/README.md image Versions are TensorFlow 2.3 and TensorFlow 1.15.

Actually, I am new to Vitis-ai, you could ask on their support forum. https://support.xilinx.com/s/?language=en_US

And I find this: image https://github.com/Xilinx/Vitis-AI/blob/1.4/docs/learn/system_requirements.md