Upwork is hiring a Stable Diffusion WebApp using Runpod.io / Baseten.co platforms

Please contact me if you're available to help us build this webapp and have immediate availability.

1. To create a webapp to allow client to be able to use these three API endpoints,

• RunPod's Stable Diffusion: https://docs.runpod.io/reference/stable-diffusion-v2

• Blueprint's Stable Diffusion: https://docs.blueprint.baseten.co/finetuning/stable-diffusion/

• Baseten's Stable Diffusion: https://app.baseten.co/explore/stable_diffusion

2. Webapp's user interface should be similar to OpenArt.ai (https://openart.ai/create?ai_model=stabilityai%2Fstable-diffusion-2-1) which will allow the user to select 1 api endpoint at a time to generate AI art

3. For each API request, webapp should save all technical information used when generating each image whenever an image is downloaded : Prompt, Model, size, scale, steps, sampler, seed, strength, creation time

4. Allow user to queue a larger number of images to be processed by the model in batches. ex: Queue 100 images, to be generated in batches of 7 at a time, and save to specific album. Allow users to queue more than one process in advance, before the first process concludes.

• One api request can generate a maximum of 2 images unlike runpod.io as they have the parameter called num_outputs to generate more than just 2 images, so either we allow user to create multiple requests and we can program the web app to execute 1 request at a time and once 1 request is completed then the next request will start progressing like a FIFO (first in - first out) sequence.

5. Allow user to send all/specific images/batches of images/specific albums to external storage (Digital Ocean Server/Google drive). Easily connect the app to various external cloud storage options

• With each submitted request, user should be able to choose between where they would like to store the generated image.

• For now we need to save the images in our server be it S3 bucket or something else, so it would be pretty straight forward.

6. Select and queue specific albums to be used for retraining more specific models. www.blueprint.baseten.co is commonly used for finetuning and deploying models.

• Blue print's fine tunning offers a few additional feature, like defining/limiting the image formats to be generated by the API. Based on our business objective, we can create our own finetuned modal on top of Blueprint and use that in the webapp.

• Ref. Link 1: https://docs.blueprint.baseten.co/dataset/stable-diffusion/

• Ref. Link 2: https://docs.blueprint.baseten.co/finetuning/stable-diffusion/

• Ref. Link 3: https://docs.blueprint.baseten.co/deployment/