https://github.com/amperecomputingai/aio-examples

https://github.com/amperecomputingai/aio-examples

Last synced: about 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/amperecomputingai/aio-examples
• Owner: AmpereComputingAI
• Created: 2021-11-11T05:53:35.000Z (about 3 years ago)
• Default Branch: tensorflow
• Last Pushed: 2024-03-05T12:33:47.000Z (10 months ago)
• Last Synced: 2024-03-05T13:47:33.787Z (10 months ago)
• Language: Python
• Size: 5.91 MB
• Stars: 2
• Watchers: 3
• Forks: 3
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# aio-examples

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.**

## Running Jupyter Notebook QuickStart Examples

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  [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.

## Running Examples With CLI

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

```