Recent ML Projects
Rebooting, again
Contents
2 months ago, I was halfway into publishing a series on my TTS project. Today, the site for that project is defunct, and no new posts have been made about it.
What happened?
Well1, I’ve been working on a lot of different projects. Some of them already exist on my Github, but only as code, without elaboration or presentation. So I’m going to dedicate the next few weeks of my time to talking about them on this space.
But before I get to writing those longer posts, here’s a quick summary of what I’ve been working on:
| Project | Status | Description | Publishing |
|---|---|---|---|
| Stable Diffusion related | ✅ | Embeddings, Dreambooth, Hypernetworks, etc | ❌ |
| Custom object detection model | Dataset done | Applying object detection models like YOLOv* to custom made datasets | ❔ |
| voice2img for Stable Diffusion | Prototype published | Using Automatic Speech Recognition (ASR) models like Whisper to generate images by vocal commands | ✅ |
| Copilot@home | ❌ | Open source alternative plugin for Copilot. Cancelled because a much better effort already exists | ✅ |
| GPT-J text editor | ✅ in progress | A webapp text editor that uses GPT-J for autocomplete | ✅ |
I’ve also had other ideas that I’ve yet to write any code for, like:
| Idea | Description | Likelihood |
|---|---|---|
| Better TTS models | Disco Narrator but actually good | 0.01 |
| More uses of Whisper | Adding to GPT-J text editor (wasm? API?). Using as a phone app | Maybe |
| Using YOLO at home | Mostly to notify for presence of visitors | Maybe |
| Research thing | Related to language models | Probably |
- Also, my computer died. I have a blog post about it I’ve yet to publish too.