GPT-J Chat
Talk directly to GPT-J model

This app logs you in to the command line of a bash shell where you can talk directly to the first large open source model, GPT-J from Eleutherai. It is completely unfiltered... but also has no fine-tuning, although most people might prefer it with a little fine-tuning. This is raw.

Web Whisper
Transcribe with OpenAI Whisper

This app gives you the ability to transcribe audio to text using the remarkably accurate Whisper Speech-To-Text (STT) engine from OpenAI. Whishper is an Open Source frontend to Whisper. It is a mature version of the 'web-whisper' app which was previously available on this lab site.

  1. Upload a video (or paste a URL for any video [or audio] file).
  2. Select which Whisper engine to use (small to large).
  3. Select the language (from quite a few).
  4. Start the transcription.

Download completed text in multiple different file formats, (from text to captions and more). For this server, we're using a single nVidia GPU to make this happen more quickly.


Stable Diffusion I
Stable Diffusion (All features)

This app allows you to play with Stable Diffusion (AI-generated images) using the incredible interface from AUTOMATIC111, which you will see is quite impressive. Their web UI sets a high standard and is so sophisticated you will benefit from using this helpful guide.

Stable Diffusion II
Stable Diffusion (Easy)

This app allows you to play with Stable Diffusion (AI-generated images) with a simple, easy interface more appropriate for mobile devices. Or anyone who just wants to simply generate an image without all the bells and whistles.

Go to Stable Diffusion (Easy)