Spaces:
Running
Running
README: note Modal usage (Best Use of Modal sponsor prize)
Browse files
README.md
CHANGED
|
@@ -43,6 +43,15 @@ B judge panel · 1–5 sparring bots · 7 the Masher · 8 the Executioner ·
|
|
| 43 |
*(This Space is a `gradio.Server` app — Gradio's engine serves the custom canvas at `/`,
|
| 44 |
and the `trace_digest` endpoint runs through Gradio's queue, callable with `gradio_client`.)*
|
| 45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
## 🎖️ Badge artifacts (everything is public)
|
| 47 |
|
| 48 |
| Badge | Evidence |
|
|
|
|
| 43 |
*(This Space is a `gradio.Server` app — Gradio's engine serves the custom canvas at `/`,
|
| 44 |
and the `trace_digest` endpoint runs through Gradio's queue, callable with `gradio_client`.)*
|
| 45 |
|
| 46 |
+
## ⚡ Built with Modal
|
| 47 |
+
|
| 48 |
+
The entire model program ran on **[Modal](https://modal.com)**: all six fine-tune
|
| 49 |
+
generations (TRL behavior-cloning on A100s), every evaluation and the
|
| 50 |
+
classifier-free-guidance sweep, the GGUF conversion (llama.cpp), and the vLLM
|
| 51 |
+
CFG inference gateway. The published fine-tune that ships in this app was trained,
|
| 52 |
+
evaluated, and converted end-to-end on Modal — comfortably inside the $250 credit
|
| 53 |
+
grant. Recipe + reproduce commands: [model card](https://huggingface.co/Jainamshahhh/parry-tactician-1.5b-merged).
|
| 54 |
+
|
| 55 |
## 🎖️ Badge artifacts (everything is public)
|
| 56 |
|
| 57 |
| Badge | Evidence |
|