Fasnettac Paper

🚀 Asteroid model Samuele Cornell/FasNetTAC_TACDataset_separatenoisy

This is an audio model trained on the TACDataset, which can be used for audio separation tasks.

🚀 Quick Start

This model was imported from Zenodo.

✨ Features

• Tags: asteroid, audio, FasNet - TAC, audio - to - audio, multichannel, beamforming
• Datasets: TACDataset, sep_noisy

📦 Installation

No installation steps provided in the original document, so this section is skipped.

💻 Usage Examples

No code examples provided in the original document, so this section is skipped.

📚 Documentation

Description

This model was trained by popcornell using the TAC/TAC recipe in Asteroid. It was trained on the separate_noisy task of the TACDataset dataset.

Training config

data:
    dev_json: ./data/validation.json
    sample_rate: 16000
    segment: None
    test_json: ./data/test.json
    train_json: ./data/train.json
net:
    chunk_size: 50
    context_ms: 16
    enc_dim: 64
    feature_dim: 64
    hidden_dim: 128
    hop_size: 25
    n_layers: 4
    n_src: 2
    window_ms: 4
optim:
    lr: 0.001
    weight_decay: 1e-06
training:
    accumulate_batches: 1
    batch_size: 8
    early_stop: True
    epochs: 200
    gradient_clipping: 5
    half_lr: True
    num_workers: 8
    patience: 30
    save_top_k: 10

Results

si_sdr: 10.871864315894744
si_sdr_imp: 11.322284052560262

🔧 Technical Details

No specific technical implementation details (>50 words) provided in the original document, so this section is skipped.

📄 License

This work "FasNetTAC_TACDataset_separatenoisy" is a derivative of LibriSpeech ASR corpus by Vassil Panayotov, used under CC BY 4.0; of End - to - end Microphone Permutation and Number Invariant Multi - channel Speech Separation by Yi Luo, Zhuo Chen, Nima Mesgarani, Takuya Yoshioka, used under CC BY 4.0. "FasNetTAC_TACDataset_separatenoisy" is licensed under Attribution - ShareAlike 3.0 Unported by popcornell.