docs: update validation results — Green 75%, White 90%, overall 92.3%
Browse files
README.md
CHANGED
|
@@ -63,16 +63,12 @@ Evaluated on 78 labeled songs from `staged_raw_material` using 30s/5s-stride chu
|
|
| 63 |
| Red | 11 | 12 | 91.7% |
|
| 64 |
| Orange | 4 | 4 | 100.0% |
|
| 65 |
| Yellow | 10 | 10 | 100.0% |
|
| 66 |
-
| Green |
|
| 67 |
| Blue | 11 | 11 | 100.0% |
|
| 68 |
| Indigo | 10 | 11 | 90.9% |
|
| 69 |
| Violet | 11 | 12 | 91.7% |
|
| 70 |
-
| White |
|
| 71 |
-
| **Overall** | **
|
| 72 |
-
|
| 73 |
-
**Green (0%)** — all predicted as Yellow. This is pipeline-safe: Green and Yellow share identical CHROMATIC_TARGETS distributions, so downstream chromatic match and drift scores are unaffected.
|
| 74 |
-
|
| 75 |
-
**White (0%)** — all predicted as Yellow or Blue. White's uniform `[0.33, 0.34, 0.33]` targets are meaningfully different, so this is a known open issue. White albums are musically intentionally diverse, which makes them acoustically diffuse in CLAP's feature space.
|
| 76 |
|
| 77 |
## Usage
|
| 78 |
|
|
|
|
| 63 |
| Red | 11 | 12 | 91.7% |
|
| 64 |
| Orange | 4 | 4 | 100.0% |
|
| 65 |
| Yellow | 10 | 10 | 100.0% |
|
| 66 |
+
| Green | 6 | 8 | 75.0% |
|
| 67 |
| Blue | 11 | 11 | 100.0% |
|
| 68 |
| Indigo | 10 | 11 | 90.9% |
|
| 69 |
| Violet | 11 | 12 | 91.7% |
|
| 70 |
+
| White | 9 | 10 | 90.0% |
|
| 71 |
+
| **Overall** | **72** | **78** | **92.3%** |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
|
| 73 |
## Usage
|
| 74 |
|