Make llama.cpp ZeroGPU backend default

Co-authored-by: Codex <chatgpt-codex-connector[bot]@users.noreply.github.com>

DEPLOY.md

@@ -40,7 +40,7 @@ The badge-target model path is the official OpenBMB GGUF repo running through ll
 LLAMACPP_GGUF_REPO=openbmb/MiniCPM-V-4.6-gguf
 LLAMACPP_MODEL_FILE=MiniCPM-V-4_6-Q4_K_M.gguf
 LLAMACPP_MMPROJ_FILE=mmproj-model-f16.gguf
-EXTRACTOR_BACKEND=llamacpp-gpu
+EXTRACTOR_BACKEND=auto
 ```
 
 The backend lives in:
@@ -102,7 +102,7 @@ Local development can run the same backend, although the model may be slow or to
 
 ```bash
 pip install -r requirements.txt
-EXTRACTOR_BACKEND=llamacpp-gpu python app.py
+EXTRACTOR_BACKEND=auto python app.py
 ```
 
 For quick UI-only work, continue using the static reference report without triggering extraction.

DEPLOYMENT_LOG.md

@@ -15,7 +15,7 @@ What changed:
 - `README.md` metadata changed from `sdk: docker` to `sdk: gradio`.
 - `DEPLOY.md` and `RUNBOOK.md` now describe Gradio + ZeroGPU + Transformers as the active path.
 - `src/extraction/zerogpu_transformers.py` adds the official OpenBMB MiniCPM-V Transformers backend.
-- `src/extraction/factory.py` now resolves `auto` to the ZeroGPU backend.
+- `src/extraction/factory.py` initially resolved `auto` to the ZeroGPU Transformers backend.
 - Docker-only files (`Dockerfile`, `start.sh`, `.dockerignore`) were removed from the active deployment.
 
 Current model:
@@ -37,7 +37,7 @@ Do not reintroduce Docker or `llama-server` while the project is targeting ZeroG
 ## 2026-06-11 — Add llama.cpp badge path on ZeroGPU
 
 Decision: keep the Space as **Gradio ZeroGPU**, but target the hackathon llama.cpp badge with the
-`llamacpp-gpu` backend.
+`auto` / `llamacpp-gpu` backend.
 
 Why:
 
@@ -47,10 +47,10 @@ Why:
 - The official OpenBMB GGUF repo stays outside the Space git repo and is downloaded through the
   Hugging Face cache at runtime.
 
-Current badge-target variables:
+Current badge-target defaults:
 
 ```bash
-EXTRACTOR_BACKEND=llamacpp-gpu
+EXTRACTOR_BACKEND=auto
 LLAMACPP_GGUF_REPO=openbmb/MiniCPM-V-4.6-gguf
 LLAMACPP_MODEL_FILE=MiniCPM-V-4_6-Q4_K_M.gguf
 LLAMACPP_MMPROJ_FILE=mmproj-model-f16.gguf

README.md

@@ -56,7 +56,7 @@ This workflow should not be further changed back to Docker unless the project in
 ZeroGPU. When the fine-tuned GGUF model is ready, only replace the model variables:
 
 ```bash
-EXTRACTOR_BACKEND=llamacpp-gpu
+EXTRACTOR_BACKEND=auto
 LLAMACPP_GGUF_REPO=openbmb/MiniCPM-V-4.6-gguf
 LLAMACPP_MODEL_FILE=MiniCPM-V-4_6-Q4_K_M.gguf
 LLAMACPP_MMPROJ_FILE=mmproj-model-f16.gguf
@@ -71,5 +71,5 @@ and only insert the fine-tuned GGUF repository/path into the `LLAMACPP_*` variab
 
 ```bash
 pip install -r requirements.txt
-EXTRACTOR_BACKEND=llamacpp-gpu python app.py
+EXTRACTOR_BACKEND=auto python app.py
 ```

RUNBOOK.md

@@ -11,7 +11,7 @@ This replaced the Docker + `llama-server` path because ZeroGPU is only available
 | Space SDK | `gradio` |
 | Hardware | ZeroGPU |
 | Badge-target runtime | `llama.cpp` through `llama-cpp-python` |
-| Badge-target backend | `EXTRACTOR_BACKEND=llamacpp-gpu` |
+| Badge-target backend | `EXTRACTOR_BACKEND=auto` or `EXTRACTOR_BACKEND=llamacpp-gpu` |
 | Fallback backend | `EXTRACTOR_BACKEND=zerogpu` with Transformers |
 | Model variables | `LLAMACPP_GGUF_REPO`, `LLAMACPP_MODEL_FILE`, `LLAMACPP_MMPROJ_FILE` |
 | Extraction backends | `src/extraction/llamacpp_gpu.py`, `src/extraction/zerogpu_transformers.py` |
@@ -23,8 +23,8 @@ Do not switch the Space back to Docker unless the project intentionally gives up
 
 `EXTRACTOR_BACKEND`:
 
-- `auto`: default, uses ZeroGPU Transformers as the safest startup path.
-- `llamacpp-gpu`: badge-target path, runs GGUF through `llama.cpp` inside `@spaces.GPU`.
+- `auto`: default badge-target path, runs GGUF through `llama.cpp` inside `@spaces.GPU`.
+- `llamacpp-gpu`: explicit alias for the same badge-target path.
 - `zerogpu`: force the ZeroGPU Transformers fallback backend.
 - `api`: hosted OpenBMB endpoint for development fallback only.
 - `local` / `server` / `llamacpp`: local experimental backends, not the active HF Space path.
@@ -63,10 +63,10 @@ Both ZeroGPU backends use `@spaces.GPU(duration=120)` for the model generation c
 
 ## Current Model
 
-Badge-target Space variables:
+Badge-target defaults:
 
 ```bash
-EXTRACTOR_BACKEND=llamacpp-gpu
+EXTRACTOR_BACKEND=auto
 LLAMACPP_GGUF_REPO=openbmb/MiniCPM-V-4.6-gguf
 LLAMACPP_MODEL_FILE=MiniCPM-V-4_6-Q4_K_M.gguf
 LLAMACPP_MMPROJ_FILE=mmproj-model-f16.gguf
@@ -110,7 +110,7 @@ For local extraction testing with the same backend:
 
 ```bash
 pip install -r requirements.txt
-EXTRACTOR_BACKEND=zerogpu python app.py
+EXTRACTOR_BACKEND=auto python app.py
 ```
 
 Local machines without a suitable GPU may be slow or may not have enough memory for full model inference. In that case, test UI/report rendering locally and test extraction on the HF ZeroGPU Space.

src/extraction/factory.py

@@ -1,8 +1,8 @@
 """Backend selection.
 
 `EXTRACTOR_BACKEND` env:
-  - `auto` / `zerogpu` (default): HF ZeroGPU + official OpenBMB Transformers model.
-  - `llamacpp-gpu` / `llama-champion`: HF ZeroGPU + llama.cpp GGUF badge path.
+  - `auto` / `llamacpp-gpu` / `llama-champion`: HF ZeroGPU + llama.cpp GGUF badge path.
+  - `zerogpu` / `transformers`: HF ZeroGPU + official OpenBMB Transformers fallback.
   - `api`: hosted OpenBMB endpoint (dev fallback only).
   - `local` / `server`: local llama-server backend for local development.
   - `llamacpp`: in-process llama-cpp-python backend for local development.
@@ -27,11 +27,11 @@ def build_extractor(
 ) -> Extractor:
     backend = os.getenv("EXTRACTOR_BACKEND", "auto").strip().lower()
 
-    if backend in ("auto", "zerogpu", "zero-gpu", "transformers"):
-        return ZeroGPUTransformersExtractor(model_id=model)
-    if backend in ("llamacpp-gpu", "gpu-llamacpp", "llama-champion"):
+    if backend in ("auto", "llamacpp-gpu", "gpu-llamacpp", "llama-champion"):
         # llama.cpp on the ZeroGPU GPU -> earns the Llama Champion badge while staying off-grid.
         return LlamaCppGPUExtractor()
+    if backend in ("zerogpu", "zero-gpu", "transformers"):
+        return ZeroGPUTransformersExtractor(model_id=model)
     if backend == "api":
         return OpenBMBExtractor(api_url=api_url, model=model, api_key=api_key)
     if backend in ("local", "server", "local-server"):