pranav commited on
Commit ·
d40e8dc
1
Parent(s): 7c7d73c
README
Browse files
README.md
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Saved Features for Algonauts '25 Dataset
|
| 2 |
+
|
| 3 |
+
This repository contains pre-extracted features for the Algonauts Challenge dataset using baseline models.
|
| 4 |
+
|
| 5 |
+
## Features Overview
|
| 6 |
+
|
| 7 |
+
The `developer_kit` directory contains features extracted for the entire dataset using the following models:
|
| 8 |
+
|
| 9 |
+
### Video Features
|
| 10 |
+
- Model: SlowFast R50
|
| 11 |
+
- Extracts spatiotemporal features from video frames
|
| 12 |
+
- Captures motion and appearance information
|
| 13 |
+
|
| 14 |
+
### Audio Features
|
| 15 |
+
- Model: MFCC (Mel-frequency cepstral coefficients)
|
| 16 |
+
- Standard acoustic features representing the short-term power spectrum
|
| 17 |
+
- Captures key audio characteristics
|
| 18 |
+
|
| 19 |
+
### Language Features
|
| 20 |
+
- Model: BERT (Bidirectional Encoder Representations from Transformers)
|
| 21 |
+
- Contextual word embeddings from text descriptions
|
| 22 |
+
- Encodes semantic meaning of language
|
| 23 |
+
|
| 24 |
+
|