John Ho

this allows us to resume training from a previous checkpoint

resume from checkpoint when calling `trainer.train()`

1

any plans to expose this to `VisionClassifierTrainer` so one can resume training from last checkpoint?

I have a ~200k images dataset so the augmentation step in `VisionDataset.fromImageFolder` will easily use up all the RAM when creating the `new_ds` I currently just set `augmentation=False` for this...

I'm on `torchvision==0.16.2` and I'm getting the following deprecation warning: ``` /usr/local/lib/python3.9/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights'...

this closes #2; the API is now accessible at https://etf4u.herokuapp.com/ to query any given etf, simply do https://etf4u.herokuapp.com/query/{etf_ticker}, for example see the request for [SPY](https://etf4u.herokuapp.com/query/spy) the API endpoint is a...

docker image/ api

1

this is a great project! any plans to create some sort of API/ Docker Image so this can be called as a micro-service?

end of Heroku free plan

10

any plans to redeploy this somewhere else now that Heroku's free plan is gone? Big fan of fillmurray here!

any plans to use a free api like wttr.in?

module diffusers has no attribute DifFacePipeline

2

followed the example code from the [huggingface model card](https://huggingface.co/OAOA/DifFace) but got the error then tried the below with the same error ``` from diffusers import DiffusionPipeline model_id = "OAOA/DifFace" #...

1. the script is missing `import numpy as np` at the top 2. seems like the code is tested on images that are 224x224? I bumped into a lot of...