| | Show HN: MaximusLLM – Train 262k-vocab LLMs on a single 16GB GPU (github.com/yousef-rafat) |
| 2 points by yousef_g 41 days ago | past |
|
| | Ghost Logits: Simulating missing partition mass in sampled softmax [pdf] (github.com/yousef-rafat) |
| 1 point by yousef_g 42 days ago | past |
|
| | Show HN: MaximusLLM, Breaking transformer's O(N^2) and O(V) scaling bottlenecks (github.com/yousef-rafat) |
| 1 point by yousef_g 44 days ago | past |
|
| | MaximusLLM: High-Speed Architecture via Ghost Logits and Random Latent Attention (github.com/yousef-rafat) |
| 1 point by yousef_g 45 days ago | past |
|
| | I have reimplemented Stable Diffusion 3.5 from scratch in pure PyTorch (github.com/yousef-rafat) |
| 481 points by yousef_g 10 months ago | past | 77 comments |
|
| | Magna: Embedding similarity search tool for searching within large documents (github.com/yousef-rafat) |
| 14 points by yousef_g on Jan 5, 2025 | past |
|
| | RustyChat: Asynchronous local chat server written in Rust (github.com/yousef-rafat) |
| 2 points by yousef_g on Dec 31, 2024 | past |
|
| | Open Source Twitter Bot (github.com/yousef-rafat) |
| 4 points by yousef_g on Sept 15, 2024 | past | 1 comment |
|
