how to use finetuner in a cloud environment

[vahuja4]

So, I ran this

`import finetuner

finetuner.fit(
model=model,
interactive=True,
train_data=data_gen,
freeze=True,
to_embedding_model=True,
input_size=(3, 224, 224),
output_dim=100
)

`
And, it completed in 20s, but the labeling UI didn't pop up because I am in a cloud based jupyter environment. What is the best way to use a cloud gpu with finetuner?

[bwanglzu]

@vahuja4 i see your issue. in notebook/cloud machines we might don't have the browser module. We'll see if it is possible to trigger a local browser for that, thanks for reporting this.

My suggestion is, for now, either:
use tailor + tuner api
or:
use it locally.

Sorry for the inconvenience.

[vahuja4]

Running it locally (on an M1 Mac) gave me this error: https://ibb.co/nM7TmMT

[vahuja4]

@tadejsv - I just used the celebA example code for my own dataset. No other changes. Used pretrained resnet50 from torchvision.

[tadejsv]

What format are images in your dataset?

[tadejsv]

If you can, paste the entire python file here - often errors appear in places you don't expect them to

[bwanglzu]

and please do verify your data path: /Users/jina/Downloads/img_align_celeba/*.jpg', the data path of the example should be replaced with your local data path.

[vahuja4]

Here is the code:

from jina.types.document.generators import from_files

def data_gen():
    for d in from_files('/Users/vishal/Downloads/earrings/*.jpg', size=100, to_dataturi=True):
        d.convert_image_datauri_to_blob(color_axis=0)  # `color_axis=-1` for TF/Keras users
        yield d
import torchvision
model = torchvision.models.resnet50(pretrained=True)
import finetuner

finetuner.fit(
    model=model,
    interactive=True,
    train_data=data_gen,
    freeze=True,
    to_embedding_model=True,
    input_size=(3, 224, 224),
    output_dim=100
)

[vahuja4]

jpg

…

On Thu, 21 Oct 2021, 19:43 Tadej Svetina, ***@***.***> wrote: What format are images in your dataset? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#155 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABILQW2QM5B7N77FL6XWV7DUIANZFANCNFSM5GNT6JWA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[bwanglzu]

@vahuja4 can you update your jpg, it's now shown up