Name classify provider classes by training scope. (#92)

openai, google, and tc260 stay provider classes. The Muse Image head
is public as muse-image, not meta. The freeze file still keys that
head meta_muse_image.

docs/cli.md

@@ -84,7 +84,7 @@ remove-ai-watermarks classify image.png --json
 
 The command runs the frozen photo detector (CLIP-L-ft ridge AND freeze MLP).
 Only a DEFINITELY result is reported as `ai`. On that result only, the 124-d
-provider head may return `openai`, `google`, `meta`, or `tc260`. POSSIBLY is
+provider head may return `openai`, `google`, `muse-image`, or `tc260`. POSSIBLY is
 `unknown`. Camera-like photographs are `human`. The call does not run cleanup
 and does not set `is_ai_generated`.
 

docs/module-internals.md

@@ -587,7 +587,8 @@ derived results influence detection.
 [`classify.py`](../src/remove_ai_watermarks/classify.py) is the 2026-08-31
 photo freeze: CLIP-L-ft ridge AND freeze MLP, then 124-d focal heads only on
 DEFINITELY. The public label is `ai` / `human` / `unknown`. POSSIBLY is
-`unknown`. Provider is `None` unless the label is `ai`. Tests in
+`unknown`. Provider is `openai` / `google` / `muse-image` / `tc260` / `None`.
+The freeze file keys the Muse Image head `meta_muse_image`. Tests in
 `tests/test_classify.py` pin the gate without downloads and pin that
 `identify` does not import this module.
 

docs/photo-classify-hf/README.md

@@ -43,8 +43,8 @@ The training catalog is not in this Hub repository and is not on GitHub.
 Use it when you have a photographic still and you want a pixel opinion after
 metadata is gone:
 
-- `label=ai` plus `provider=openai|google|meta|tc260` on a stripped ChatGPT,
-  Gemini, Meta Muse, or TC260 photograph
+- `label=ai` plus `provider=openai|google|muse-image|tc260` on a stripped ChatGPT,
+  Gemini, Muse Image, or TC260 photograph
 - `label=human` on camera photographs at a low false-positive rate
 - `label=unknown` when the detector is only POSSIBLY AI, or when 124-d
   features cannot be extracted
@@ -78,7 +78,7 @@ flowchart TD
   m1 -->|likely_human| human[label human, provider none]
   m1 -->|possibly| unk[label unknown, provider none]
   m1 -->|definitely| m2[Model 2: 124-d one-vs-rest focal]
-  m2 -->|openai google meta tc260| named[label ai plus provider]
+  m2 -->|openai google muse-image tc260| named[label ai plus provider]
   m2 -->|no_ai or extract fail| aiOnly[label ai, provider none]
 ```
 
@@ -89,8 +89,9 @@ remain open. The shipped heads do not include a receipt specialist.
 Provider attribution tracks the **renderer**, not the front-end. Bing Image
 Creator rows signed Microsoft, OpenAI score as `openai`. Designer rows signed
 Microsoft, Google LLC score as `google`. Native Designer is mixed. There is
-no Microsoft pixel class. Instagram `made_with_ai` is not a Meta pixel class.
-Meta requires known-provenance Muse Image output.
+no Microsoft pixel class. `openai`, `google`, and `tc260` are provider classes.
+`muse-image` is Muse Image output, not a general Meta class. Instagram
+`made_with_ai` is not that class.
 
 ## Architecture
 
@@ -115,8 +116,9 @@ Meta requires known-provenance Muse Image output.
   (`openai`, `google`, `tc260`, `meta_muse_image`, `no_ai`)
 - Decision: a class wins only if it beats `no_ai` by margin 0.30, then argmax
   among those that passed
-- Public name `meta` is the Muse Image class. `no_ai` is not reported as
-  `human`; it becomes `provider=None` on an `ai` label (for example FLUX)
+- Public name `muse-image` is the Muse Image class. The freeze file still
+  keys that head `meta_muse_image`. `no_ai` is not reported as `human`; it
+  becomes `provider=None` on an `ai` label (for example FLUX)
 
 ## Training
 

docs/photo-classify-hf/operating-point.json

@@ -21,9 +21,9 @@
   "model2": {
     "feature_width": 124,
     "margin": 0.3,
-    "classes": ["openai", "google", "tc260", "meta_muse_image", "no_ai"],
-    "public_names": {
-      "meta_muse_image": "meta"
+    "classes": ["openai", "google", "tc260", "muse-image", "no_ai"],
+    "checkpoint_keys": {
+      "muse-image": "meta_muse_image"
     },
     "runs_only_after": "definitely"
   },

docs/photo-classify-training.md

@@ -78,7 +78,8 @@ construction and fits the same heads from the pack above.
    Threshold is the 1.67% quantile on `detector_photo_dev_oi`.
 2. AND that MLP with the frozen ridge in the probe file. DEFINITELY is both.
 3. Load 124-d vectors. Fit one-vs-rest focal heads for openai, google, tc260,
-   meta_muse_image, no_ai. Margin 0.30.
+   meta_muse_image, no_ai. Margin 0.30. The public class for the Muse Image
+   head is `muse-image`.
 4. Repeat with the same seeds. Compare bytes.
 
 CLIP-L fine-tune of the last two vision blocks is earlier, seed 20260822, and

docs/photo-classify.md

@@ -54,7 +54,7 @@ flowchart TD
   m1 -->|likely_human| human["label human, provider none"]
   m1 -->|possibly| unk["label unknown, provider none"]
   m1 -->|definitely| m2[Model 2 124-d focal heads]
-  m2 -->|openai google meta tc260| named["label ai plus provider"]
+  m2 -->|openai google muse-image tc260| named["label ai plus provider"]
   m2 -->|no_ai or extract fail| aiOnly["label ai, provider none"]
 ```
 
@@ -63,7 +63,7 @@ flowchart TD
 | `label` | `ai`, `human`, `unknown` | Public verdict. `ai` only on DEFINITELY |
 | `domain` | `photo` | This freeze is photographic only |
 | `detector` | `definitely`, `possibly`, `likely_human` | Raw Model 1 gate |
-| `provider` | `openai`, `google`, `meta`, `tc260`, or `None` | Model 2, only if `label` is `ai` |
+| `provider` | `openai`, `google`, `muse-image`, `tc260`, or `None` | Model 2, only if `label` is `ai` |
 
 `human` is camera-like under this contract, not a proof of authorship.
 `unknown` is POSSIBLY, or a domain this freeze does not support. `provider=None`
@@ -84,8 +84,9 @@ wrong job on that bank, which is why Model 2 is gated.
 
 Provider names the renderer, not the product UI. Bing Image Creator signed
 Microsoft, OpenAI scores `openai`. Designer signed Microsoft, Google LLC
-scores `google`. There is no Microsoft pixel class. Meta is Muse Image, not
-an Instagram `made_with_ai` tag.
+scores `google`. There is no Microsoft pixel class. `openai`, `google`, and
+`tc260` are provider classes. `muse-image` is Muse Image output, not a
+general Meta class and not an Instagram `made_with_ai` tag.
 
 ## Evaluation
 

docs/python-api.md

@@ -160,7 +160,7 @@ print(result.label, result.detector, result.provider)
 
 `label` is `ai` only on a DEFINITELY detector result (ridge AND freeze MLP).
 POSSIBLY is `unknown`. Camera-like photographs are `human`. `provider` is
-`openai`, `google`, `meta`, or `tc260` only when `label` is `ai` and the 124-d
+`openai`, `google`, `muse-image`, or `tc260` only when `label` is `ai` and the 124-d
 head beats `no_ai` by the freeze margin. Otherwise it is `None`, including
 when 124-d extraction refuses the file.
 

src/remove_ai_watermarks/classify.py

@@ -35,17 +35,19 @@ MLP_THRESHOLD = 5.9586493237495395
 RIDGE_THRESHOLD = 0.3056212276800537
 PROVIDER_MARGIN = 0.30
 CLIP_WIDTH = 768
+# Checkpoint keys in provider.pt. openai/google/tc260 are provider classes;
+# meta_muse_image is Muse Image output. Public names follow that split.
 PROVIDER_LABELS = ("openai", "google", "tc260", "meta_muse_image", "no_ai")
 
 DetectorLevel = Literal["definitely", "possibly", "likely_human"]
 PixelLabel = Literal["ai", "human", "unknown"]
 PixelDomain = Literal["photo"]
-PixelProvider = Literal["openai", "google", "meta", "tc260"]
+PixelProvider = Literal["openai", "google", "muse-image", "tc260"]
 PUBLIC_PROVIDER: dict[str, PixelProvider] = {
     "openai": "openai",
     "google": "google",
     "tc260": "tc260",
-    "meta_muse_image": "meta",
+    "meta_muse_image": "muse-image",
 }
 
 _runtime: _Runtime | None = None
@@ -104,7 +106,7 @@ def label_for(level: DetectorLevel) -> PixelLabel:
 
 
 def provider_from_scores(scores: dict[str, float], *, margin: float = PROVIDER_MARGIN) -> PixelProvider | None:
-    """Argmax among openai/google/tc260/meta that beat ``no_ai`` by ``margin``."""
+    """Argmax among openai/google/tc260/muse-image that beat ``no_ai`` by ``margin``."""
     no_ai = scores["no_ai"]
     ai_names = [name for name in PROVIDER_LABELS if name != "no_ai"]
     passed = [name for name in ai_names if scores[name] > no_ai + margin]

tests/test_classify.py

@@ -55,6 +55,8 @@ def test_shipped_operating_point_matches_the_runtime_defaults() -> None:
     assert payload["model1"]["ridge_threshold"] == RIDGE_THRESHOLD
     assert payload["model2"]["margin"] == PROVIDER_MARGIN
     assert payload["model2"]["runs_only_after"] == "definitely"
+    assert payload["model2"]["classes"] == ["openai", "google", "tc260", "muse-image", "no_ai"]
+    assert payload["model2"]["checkpoint_keys"]["muse-image"] == "meta_muse_image"
 
 
 def test_definitely_is_the_and_of_ridge_and_mlp() -> None:
@@ -74,7 +76,7 @@ def test_provider_requires_margin_over_no_ai() -> None:
     assert provider_from_scores(_scores(openai=1.0, no_ai=0.5)) == "openai"
     assert provider_from_scores(_scores(openai=0.5, no_ai=0.4)) is None
     assert provider_from_scores(_scores(google=2.0, openai=1.5, no_ai=0.0)) == "google"
-    assert provider_from_scores(_scores(meta=1.0, no_ai=0.0)) == "meta"
+    assert provider_from_scores(_scores(meta=1.0, no_ai=0.0)) == "muse-image"
     assert provider_from_scores(_scores(tc260=1.0, no_ai=0.0)) == "tc260"