Getting started

Installation

We recommend using Miniconda to manage your python versions and environments.

Step 1: Python environment

Install miniconda on your machine.

Windows: When the installation finishes, open Anaconda Prompt from the Start menu.

macOS/Linux: When the installation finished, open a new shell window.

Set up an environment for Qualcomm® AI Hub:

conda create python=3.8 -n qai_hub
conda activate qai_hub

Warning

If any of these steps fail with SSL: CERTIFICATE_VERIFY_FAILED errors, you have an SSL interception and traffic inspection tool installed. Ask your IT department for instructions on how to set up certificates for Python pip and the Python Requests library. They should provide you with a new certificate file. You will also need to add this certificate to Miniconda prior to creating your environment. This can be done with conda config --set ssl_verify <path to certificate>.

Step 2: Install Python client

Note

Windows for ARM: AI Hub is currently not supported using native ARM Python due to lack of support for depenendencies. While the Python ecosystem catches up to ARM Windows, please install the x64 version of Python (made possible through emulation). We recommend Python 3.8 “64-bit” from the Python download page.

pip3 install qai-hub

Step 3: Sign in

Go to Qualcomm® AI Hub and sign in with your Qualcomm ID to view information about jobs you create.

Once signed in navigate to Account -> Settings -> API Token. This should provide an API token that you can use to configure your client.

Step 4: Configure API Token

Next, configure the client with API token using the following command in your terminal:

qai-hub configure --api_token INSERT_API_TOKEN

You can check that your API token is installed correctly by fetching a list of available devices. To do that, you can type the following in a Python terminal:

import qai_hub as hub
hub.get_devices()

You’re now set up to submit jobs on Qualcomm AI Hub. If you run into any issues, please contact us. Feel free try out the examples below to run a profile job on the desired device.

Quick example (PyTorch)

Once you have set up your Qualcomm® AI Hub environment, the next step is to submit a profiling job. First, install the dependencies of this example:

pip3 install "qai-hub[torch]"

Warning

If any of the snippets fail with an API authentication error, it means that you do not have a valid API token installed. Please see Installation to learn how to set this up.

Submit an analysis of the MobileNet v2 network on the device type of your choice. You can use any device from the list above to specify a different device than the options below.

Mobile

import numpy as np
import requests
import torch
from PIL import Image
from torchvision.models import mobilenet_v2

import qai_hub as hub

# Using pre-trained MobileNet
torch_model = mobilenet_v2(pretrained=True)
torch_model.eval()

# Step 1: Trace model
input_shape = (1, 3, 224, 224)
example_input = torch.rand(input_shape)
traced_torch_model = torch.jit.trace(torch_model, example_input)

# Step 2: Compile model
compile_job = hub.submit_compile_job(
    model=traced_torch_model,
    device=hub.Device("Samsung Galaxy S24 (Family)"),
    input_specs=dict(image=input_shape),
    options="--target_runtime tflite",
)
assert isinstance(compile_job, hub.CompileJob)

# Step 3: Profile on cloud-hosted device
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
profile_job = hub.submit_profile_job(
    model=target_model,
    device=hub.Device("Samsung Galaxy S24 (Family)"),
)
assert isinstance(profile_job, hub.ProfileJob)


# Step 4: Run inference on cloud-hosted device
sample_image_url = (
    "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/input_image1.jpg"
)
response = requests.get(sample_image_url, stream=True)
response.raw.decode_content = True
image = Image.open(response.raw).resize((224, 224))
input_array = np.expand_dims(
    np.transpose(np.array(image, dtype=np.float32) / 255.0, (2, 0, 1)), axis=0
)

# Run inference using the on-device model on the input image
inference_job = hub.submit_inference_job(
    model=target_model,
    device=hub.Device("Samsung Galaxy S24 (Family)"),
    inputs=dict(image=[input_array]),
)
assert isinstance(inference_job, hub.InferenceJob)

on_device_output = inference_job.download_output_data()
assert isinstance(on_device_output, dict)

# Step 5: Post-processing the on-device output
output_name = list(on_device_output.keys())[0]
out = on_device_output[output_name][0]
on_device_probabilities = np.exp(out) / np.sum(np.exp(out), axis=1)

# Read the class labels for imagenet
sample_classes = "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/imagenet_classes.txt"
response = requests.get(sample_classes, stream=True)
response.raw.decode_content = True
categories = [str(s.strip()) for s in response.raw]

# Print top five predictions for the on-device model
print("Top-5 On-Device predictions:")
top5_classes = np.argsort(on_device_probabilities[0], axis=0)[-5:]
for c in reversed(top5_classes):
    print(f"{c} {categories[c]:20s} {on_device_probabilities[0][c]:>6.1%}")

# Step 6: Download model
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
target_model.download("mobilenet_v2.tflite")

IOT

import numpy as np
import requests
import torch
from PIL import Image
from torchvision.models import mobilenet_v2

import qai_hub as hub

# Using pre-trained MobileNet
torch_model = mobilenet_v2(pretrained=True)
torch_model.eval()

# Step 1: Trace model
input_shape = (1, 3, 224, 224)
example_input = torch.rand(input_shape)
traced_torch_model = torch.jit.trace(torch_model, example_input)

# Step 2: Compile model
compile_job = hub.submit_compile_job(
    model=traced_torch_model,
    device=hub.Device("RB3 Gen 2 (Proxy)"),
    input_specs=dict(image=input_shape),
    options="--target_runtime tflite",
)
assert isinstance(compile_job, hub.CompileJob)


# Step 3: Profile on cloud-hosted device
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
profile_job = hub.submit_profile_job(
    model=target_model,
    device=hub.Device("RB3 Gen 2 (Proxy)"),
)
assert isinstance(profile_job, hub.ProfileJob)


# Step 4: Run inference on cloud-hosted device
sample_image_url = (
    "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/input_image1.jpg"
)
response = requests.get(sample_image_url, stream=True)
response.raw.decode_content = True
image = Image.open(response.raw).resize((224, 224))
input_array = np.expand_dims(
    np.transpose(np.array(image, dtype=np.float32) / 255.0, (2, 0, 1)), axis=0
)

# Run inference using the on-device model on the input image
inference_job = hub.submit_inference_job(
    model=target_model,
    device=hub.Device("RB3 Gen 2 (Proxy)"),
    inputs=dict(image=[input_array]),
)
assert isinstance(inference_job, hub.InferenceJob)

on_device_output = inference_job.download_output_data()
assert isinstance(on_device_output, dict)

# Step 5: Post-processing the on-device output
output_name = list(on_device_output.keys())[0]
out = on_device_output[output_name][0]
on_device_probabilities = np.exp(out) / np.sum(np.exp(out), axis=1)

# Read the class labels for imagenet
sample_classes = "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/imagenet_classes.txt"
response = requests.get(sample_classes, stream=True)
response.raw.decode_content = True
categories = [str(s.strip()) for s in response.raw]

# Print top five predictions for the on-device model
print("Top-5 On-Device predictions:")
top5_classes = np.argsort(on_device_probabilities[0], axis=0)[-5:]
for c in reversed(top5_classes):
    print(f"{c} {categories[c]:20s} {on_device_probabilities[0][c]:>6.1%}")

# Step 6: Download model
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
target_model.download("mobilenet_v2.tflite")

Compute

import numpy as np
import requests
import torch
from PIL import Image
from torchvision.models import mobilenet_v2

import qai_hub as hub

# Using pre-trained MobileNet
torch_model = mobilenet_v2(pretrained=True)
torch_model.eval()

# Step 1: Trace model
input_shape = (1, 3, 224, 224)
example_input = torch.rand(input_shape)
traced_torch_model = torch.jit.trace(torch_model, example_input)

# Step 2: Compile model
compile_job = hub.submit_compile_job(
    model=traced_torch_model,
    device=hub.Device("Snapdragon X Elite CRD"),
    input_specs=dict(image=input_shape),
    options="--target_runtime onnx",
)
assert isinstance(compile_job, hub.CompileJob)

# Step 3: Profile on cloud-hosted device
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
profile_job = hub.submit_profile_job(
    model=target_model,
    device=hub.Device("Snapdragon X Elite CRD"),
)
assert isinstance(profile_job, hub.ProfileJob)


# Step 4: Run inference on cloud-hosted device
sample_image_url = (
    "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/input_image1.jpg"
)
response = requests.get(sample_image_url, stream=True)
response.raw.decode_content = True
image = Image.open(response.raw).resize((224, 224))
input_array = np.expand_dims(
    np.transpose(np.array(image, dtype=np.float32) / 255.0, (2, 0, 1)), axis=0
)

# Run inference using the on-device model on the input image
inference_job = hub.submit_inference_job(
    model=target_model,
    device=hub.Device("Snapdragon X Elite CRD"),
    inputs=dict(image=[input_array]),
)
assert isinstance(inference_job, hub.InferenceJob)

on_device_output = inference_job.download_output_data()
assert isinstance(on_device_output, dict)

# Step 5: Post-processing the on-device output
output_name = list(on_device_output.keys())[0]
out = on_device_output[output_name][0]
on_device_probabilities = np.exp(out) / np.sum(np.exp(out), axis=1)

# Read the class labels for imagenet
sample_classes = "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/imagenet_classes.txt"
response = requests.get(sample_classes, stream=True)
response.raw.decode_content = True
categories = [str(s.strip()) for s in response.raw]

# Print top five predictions for the on-device model
print("Top-5 On-Device predictions:")
top5_classes = np.argsort(on_device_probabilities[0], axis=0)[-5:]
for c in reversed(top5_classes):
    print(f"{c} {categories[c]:20s} {on_device_probabilities[0][c]:>6.1%}")

# Step 6: Download model
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
target_model.download("mobilenet_v2.onnx")

Automotive

import numpy as np
import requests
import torch
from PIL import Image
from torchvision.models import mobilenet_v2

import qai_hub as hub

# Using pre-trained MobileNet
torch_model = mobilenet_v2(pretrained=True)
torch_model.eval()

# Step 1: Trace model
input_shape = (1, 3, 224, 224)
example_input = torch.rand(input_shape)
traced_torch_model = torch.jit.trace(torch_model, example_input)

# Step 2: Compile model
compile_job = hub.submit_compile_job(
    model=traced_torch_model,
    device=hub.Device("SA8650 (Proxy)"),
    input_specs=dict(image=input_shape),
    options="--target_runtime qnn_lib_aarch64_android",
)
assert isinstance(compile_job, hub.CompileJob)

# Step 3: Profile on cloud-hosted device
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
profile_job = hub.submit_profile_job(
    model=target_model,
    device=hub.Device("SA8650 (Proxy)"),
)
assert isinstance(profile_job, hub.ProfileJob)

# Step 4: Run inference on cloud-hosted device
sample_image_url = (
    "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/input_image1.jpg"
)
response = requests.get(sample_image_url, stream=True)
response.raw.decode_content = True
image = Image.open(response.raw).resize((224, 224))
input_array = np.expand_dims(
    np.transpose(np.array(image, dtype=np.float32) / 255.0, (2, 0, 1)), axis=0
)

# Run inference using the on-device model on the input image
inference_job = hub.submit_inference_job(
    model=target_model,
    device=hub.Device("SA8650 (Proxy)"),
    inputs=dict(image=[input_array]),
)
assert isinstance(inference_job, hub.InferenceJob)

on_device_output = inference_job.download_output_data()
assert isinstance(on_device_output, dict)

# Step 5: Post-processing the on-device output
output_name = list(on_device_output.keys())[0]
out = on_device_output[output_name][0]
on_device_probabilities = np.exp(out) / np.sum(np.exp(out), axis=1)

# Read the class labels for imagenet
sample_classes = "https://qaihub-public-assets.s3.us-west-2.amazonaws.com/apidoc/imagenet_classes.txt"
response = requests.get(sample_classes, stream=True)
response.raw.decode_content = True
categories = [str(s.strip()) for s in response.raw]

# Print top five predictions for the on-device model
print("Top-5 On-Device predictions:")
top5_classes = np.argsort(on_device_probabilities[0], axis=0)[-5:]
for c in reversed(top5_classes):
    print(f"{c} {categories[c]:20s} {on_device_probabilities[0][c]:>6.1%}")

# Step 6: Download model
target_model = compile_job.get_target_model()
assert isinstance(target_model, hub.Model)
target_model.download("mobilenet_v2.so")

This will submit a compile job and then a profile job, printing the URLs of both jobs. Finally, the code performs an inference job on some sample data. View all your jobs at /jobs/.

You can programmatically query the status of the job:

status = profile_job.get_status()
print(status)

You can access the results of the job using the snippet below. There are three main parts

• Profile: Results of the profiling in JSON format.

• Target Model: Optimized model ready for deployment.

• Results: Folder containing all the artifacts of the job (including logs).

Note that these are blocking API calls that wait until the job finishes:

# Download profile results as JSON (blocking call) profile = profile_job.download_profile() print(profile)

# Download an optimized model (blocking call) model = profile_job.model.download() print(model)

# Download results to current directory (blocking call) profile_job.download_results(“.”)

For more information, continue on to Profiling Models or refer to the API documentation.