Instructions to use nvlz/ravdess-emotion-recognition with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Keras
How to use nvlz/ravdess-emotion-recognition with Keras:
# Available backend options are: "jax", "torch", "tensorflow". import os os.environ["KERAS_BACKEND"] = "jax" import keras model = keras.saving.load_model("hf://nvlz/ravdess-emotion-recognition") - Notebooks
- Google Colab
- Kaggle
RAVDESS Duygu Tanıma — BiLSTM + Temporal Attention
Test Accuracy: %80.2 (RAVDESS, 8 sınıf)
Dosyalar
| Dosya | Açıklama |
|---|---|
best_lstm.keras |
Keras modeli (custom layer: TemporalAttention) |
lstm_norm.npy |
Normalizasyon parametreleri — shape (2, 80): [mean, std] |
Mimari
- Input: (282 zaman adımı × 80 özellik) — 40 MFCC + 40 Δ MFCC
- Bidirectional LSTM(128) + LayerNorm
- Temporal Attention
- Dense(128) → BatchNorm → Dense(64) → Dense(8, softmax)
Sınıflar
neutral, calm, happy, sad, angry, fearful, disgust, surprised
Demo
ravdess_demo.ipynb ile Gradio arayüzü üzerinden test edebilirsin.
- Downloads last month
- 45
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support