Org card: contributions are welcome
Browse files
README.md
CHANGED
|
@@ -25,6 +25,10 @@ compute-adaptive ("budget") tokenization. A generative model is fine-tuned so th
|
|
| 25 |
number of tokens, chosen at inference: detailed content keeps fine tokens, simple content is merged into coarse ones,
|
| 26 |
and compute and memory fall with the token count.
|
| 27 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
> **Unofficial, open-source implementation.** The method comes from
|
| 29 |
> [BudgetPix: Compute-Adaptive Tokenization for Pixel-Space Image Diffusion](https://arxiv.org/abs/2610.12307)
|
| 30 |
> (Kara et al., 2026). The code and models here are a community reimplementation and extension of it, not an
|
|
|
|
| 25 |
number of tokens, chosen at inference: detailed content keeps fine tokens, simple content is merged into coarse ones,
|
| 26 |
and compute and memory fall with the token count.
|
| 27 |
|
| 28 |
+
**Contributions are welcome.** Bug reports, questions and pull requests on [GitHub](https://github.com/BudgetAny) are
|
| 29 |
+
all appreciated, from a typo fix or clearer documentation to a port to a new base model or a new modality. For larger
|
| 30 |
+
changes, please open an issue first so we can plan it together.
|
| 31 |
+
|
| 32 |
> **Unofficial, open-source implementation.** The method comes from
|
| 33 |
> [BudgetPix: Compute-Adaptive Tokenization for Pixel-Space Image Diffusion](https://arxiv.org/abs/2610.12307)
|
| 34 |
> (Kara et al., 2026). The code and models here are a community reimplementation and extension of it, not an
|