Hugging Face Pro Version Question

[punter82]

I am right now using the Hugging face @huggingface/inference NodeJS SDK with this model and free version.. i am interested in buying the prod version which is 20K requests per day.. but what i want to know is if i purchase pro version,

1. Will the image generation be faster.. right now it is very very slow and by using pro version
2. Will the context of the image generated will be better
3. Will i be able to generate multiple images per prompt
4. can i continue using same model or use a different one..

Your response is highly appreciated

[julien-c]

what model are you using?

[punter82]

I am using black-forest-labs/FLUX.1-schnell @julien-c

[punter82]

Any update here ?

[punter82]

Hugging face team i am so inclined to buying the pro version but can someone confirm the questions i asked.. coz i am not able to clearly find these in any docs.. @julien-c

[coyotte508]

As far I understand, the main reason for the PRO subscription is increased rate limits, increased zero-gpu quota, and access to more models. (and some private storage too!)

To know if a model is supported for pro subscribers, you can make a request to it with the free inference API, and if you have an error message that you need to be a PRO subscribers, then that model is available for PRO users.

And if you want more control you can switch over to inference endpoints.

[coyotte508]

i am not able to clearly find these in any docs

cc @Michellehbn for doc clarification maybe

[punter82]

ok thanks @coyotte508 Can you suggest which is the best model hugging face provides for text to image in the pro version ?

[punter82]

So i did purchase the pro version and looks good and using black-forest-labs/FLUX.1-schnell is there any better model you can suggest it would be great

[coyotte508]

you can use the default model (the one in https://huggingface.co/tasks/text-to-image), you can also try out the models here: https://huggingface.co/models?pipeline_tag=text-to-image sorted by popularity

[punter82]

Thanks so much @coyotte508