You can try Tensorflow powered by AIO by either running jupyter notebook examples or python scripts on CLI level.
Note: Before running the examples, please run download_models.sh script to pull down all models.
Use AIO_NUM_THREADS to specify the number of cores the AIO compute kernels will run on
export AIO_NUM_THREADS=16
cd /aio-examples/
bash download_models.sh
bash start_notebook.sh
if you would like to run examples using with CLI you can run the start_notebook.sh in the background like so:
bash start_notebook.sh &
If you run it on a cloud instance, make sure your machine has port 8080 open and on your local device run:
ssh -N -L 8080:localhost:8080 -i <ssh key> [email protected]
Use a browser to point to the URL printed out by the Jupyter notebook launcher. You will find Jupyter Notebook examples, examples.ipynb, under /classification and /object_detection folders. The examples run through several inference models, visualize results and present the performance numbers.
To use CLI-level scripts:
Use AIO_NUM_THREADS to specify the number of cores the AIO compute kernels will run on
export AIO_NUM_THREADS=16
cd /aio-examples/
Download the models:
bash download_models.sh
Go to the directory of choice, e.g.
cd classification/resnet_50_v15
Evaluate the model with run.py script
Optional arguments:
-h, --help show this help message and exit
-m MODEL_PATH, --model_path MODEL_PATH
-p {fp32,fp16,int8}, --precision {fp32,fp16,int8}
-b BATCH_SIZE, --batch_size BATCH_SIZE
python run.py -m resnet_50_v15_tf_fp32.pb -p fp32
python run.py -m resnet_50_v15_tflite_int8.tflite -p int8