How do you run this and the other QualComm Hub models on the S23 Ultra please?

How do you run this on the S23 Ultra please. about ai-hub-models HOT 2 CLOSED

AntDX316 commented on May 18, 2024

How do you run this on the S23 Ultra please.

from ai-hub-models.

Comments (2)

heechan12 commented on May 18, 2024

I'm really curious too, especially about the Stable diffusion...

from ai-hub-models.

bhushan23 commented on May 18, 2024

There's a device parameter that you can provide while submitting AI Hub jobs to select devices of interest.
e.g. example in https://app.aihub.qualcomm.com/docs/hub/compile_examples.html#compile-a-torchscript-model

In order to get complete list of devices you can run following in cli

qai-hub list-devices

and then use device name or any attribute of interest to get hub device
e.g. Following is one of the option to get a device with device name

hub.Device(name="Samsung Galaxy S23 Ultra")

Let's see this in action

import qai_hub as hub

# Compile a model
compile_job = hub.submit_compile_job(
    model="mobilenet_v2.pt",
    device=hub.Device("Samsung Galaxy S23 Ultra"),
    input_specs=dict(image=(1, 3, 224, 224)),
)
assert isinstance(compile_job, hub.CompileJob)

# Profile the compiled model
profile_job = hub.submit_profile_job(
    model=compile_job.get_target_model(),
    device=hub.Device("Samsung Galaxy S23 Ultra"),
)
assert isinstance(profile_job, hub.ProfileJob)

@heechan12 demo.py does have an option to override default device and you can do so as follow

pip install "qai_hub_models[stable_diffusion_quantized]"
python -m qai_hub_models.models.stable_diffusion_quantized.demo --device-name "Samsung Galaxy S23 Ultra"

from ai-hub-models.

Recommend Projects