Spaces:
Running
Running
|
Download README.md from anilpai/srijika-demo: direct link, hf CLI and curl.
- Browser
- Download file 1.23 kB
-
https://huggingface.co/spaces/anilpai/srijika-demo/resolve/main/README.md
- Command line
-
hf download hf://spaces/anilpai/srijika-demo/README.md
-
curl -L -o README.md https://huggingface.co/spaces/anilpai/srijika-demo/resolve/main/README.md
1.23 kB
| title: Srijika — Text to Indic Font | |
| emoji: ✒️ | |
| colorFrom: yellow | |
| colorTo: red | |
| sdk: gradio | |
| sdk_version: 5.49.1 | |
| app_file: app.py | |
| pinned: false | |
| license: mit | |
| short_description: Describe a font in words — AI draws it in Indic scripts | |
| # Srijika · सृजिका — text-to-font studio | |
| Type *"a heavy rounded poster font"* and get a working **TTF** in | |
| Devanagari, Tamil, Bengali, Telugu, Kannada, Malayalam, Gujarati, Gurmukhi or Odia, drawn | |
| glyph-by-glyph by a | |
| diffusion model. | |
| **Pipeline** — [Lipika-FontCLIP](https://huggingface.co/loopdesk-ai/lipika-fontclip) | |
| embeds your description and retrieves the closest real Indic font from a | |
| 1,700-face corpus; that donor seeds Srijika, a private glyph-diffusion | |
| model that redraws the alphabet in the matched style. Parametric axes | |
| (weight / counter / em-fill) then derive a whole family from one generation. | |
| This Space is a thin client: model weights stay on the private backend, | |
| and the demo runs draft-quality (30 glyphs, ~2 min) with shared rate | |
| limits. Companion models: [loopdesk-ai/lipika](https://huggingface.co/loopdesk-ai/lipika) | |
| (font recognition), [loopdesk-ai/lipika-fontclip](https://huggingface.co/loopdesk-ai/lipika-fontclip) | |
| (text↔font retrieval). | |