omni: CI smoke + docs + README preamble

- .gitea/workflows/smoke.yml: gitea CI on ailab gpu runner (manual git
  clone since actions/checkout@v4 mis-resolves subpath gitea); injects
  WANDB_API_KEY + CI_RUN_TAG=smoke-$run_number
- scripts/smoke.sh: in-place smoke (uv sync + 1 shard + tokenizer +
  d=6 50-step base_train); idempotent cache at /data/nanochat-smoke/
- doc/research_feasibility.md: voice-first multimodal feasibility study (mochi)
- doc/todo.md: phase-by-phase roadmap (W1 Whisper smoke → W4 MVP)
- README.md: omni preamble pointing at upstream nanochat README
- .gitignore: exclude .claude/ runtime files

README.md

@@ -1,4 +1,18 @@
-# nanochat
+# nanochat-omni
+
+> **质感感知语音输入** — multimodal extension on top of [karpathy/nanochat](https://github.com/karpathy/nanochat).
+>
+> Audio-first (Whisper encoder + Projector → soft tokens, LLaVA-style alignment).
+> Vision later. Output stays text. Single GPU (RTX 5090 / 4090). CN mirrors baked
+> into `pyproject.toml` / `nanochat/dataset.py` (sjtu pytorch-wheels, hf-mirror).
+>
+> Roadmap: [`doc/todo.md`](doc/todo.md) · Research: [`doc/research_feasibility.md`](doc/research_feasibility.md) · CI smoke: [`scripts/smoke.sh`](scripts/smoke.sh)
+>
+> Sync upstream: `git fetch upstream && git merge upstream/master`
+
+---
+
+# nanochat (upstream)
 
 ![nanochat logo](dev/nanochat.png)
 ![scaling laws](dev/scaling_laws_jan26.png)