Update README.md
Browse files
README.md
CHANGED
|
@@ -40,7 +40,25 @@ Annotations are stored under `data/annotation/`:
|
|
| 40 |
- `MADB_data.csv`: full annotation data, including:
|
| 41 |
- per-annotator scores
|
| 42 |
- textual comments
|
| 43 |
-
- genre and mood tags
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
---
|
| 46 |
|
|
|
|
| 40 |
- `MADB_data.csv`: full annotation data, including:
|
| 41 |
- per-annotator scores
|
| 42 |
- textual comments
|
| 43 |
+
- genre and mood tags
|
| 44 |
+
- 'split_val_seed5.csv': the validate set with random seed 5
|
| 45 |
+
|
| 46 |
+
---
|
| 47 |
+
|
| 48 |
+
### Sample
|
| 49 |
+
|
| 50 |
+
Samples are stored under `sample/`:
|
| 51 |
+
|
| 52 |
+
- sample_audio: 200 audios selected from validate set:
|
| 53 |
+
- 100 audios randomly selected from Muchin
|
| 54 |
+
- 50 audios from levo
|
| 55 |
+
- 50 audios from suno
|
| 56 |
+
- sample_embedding: embedding extracted from sample audios:
|
| 57 |
+
- clap: extracted with original clap
|
| 58 |
+
- muq: extracted with original muq
|
| 59 |
+
- clap_com: extracted with clap after adaption with comments
|
| 60 |
+
- clap_com_tag: extracted with clap after adaption with comments and tags
|
| 61 |
+
- 'sample_ids.csv': contains all sample audios' id
|
| 62 |
|
| 63 |
---
|
| 64 |
|