NEWVectors or files. Pick a path.Start →
    Models/Speech & Audio/facebook/wav2vec2-large-960h
    HFTranscriptionapache-2.0

    wav2vec2-large-960h

    by facebook

    Self-supervised speech representations for automatic speech recognition

    32Kdl/month
    35likes
    317Mparams
    Identifiers
    Model ID
    facebook/wav2vec2-large-960h
    Feature URI
    mixpeek://transcription@v1/facebook_wav2vec2_large_v1

    Overview

    Wav2Vec 2.0 learns speech representations from raw audio through self-supervised pre-training, then fine-tunes with a small amount of labeled data. The 960h variant is fine-tuned on the full LibriSpeech dataset.

    On Mixpeek, Wav2Vec2 provides an alternative to Whisper for English transcription, with strong performance on clear speech and a smaller memory footprint.

    Architecture

    CNN feature encoder (7 convolutional layers) followed by a 24-layer Transformer. Self-supervised pre-training uses contrastive loss over quantized speech representations. Fine-tuned with CTC loss.

    Mixpeek SDK Integration

    // No extractor parameter takes a Hugging Face model id (checked against
    // GET /v1/discovery/extractors, which returns 13), so wav2vec2-large-960h runs
    // on your side and the output is upserted through POST
    // /v1/namespaces/{namespace_id}/documents/upsert. On Enterprise the other
    // path is to upload the weights instead: POST /v1/namespaces/{id}/models
    // accepts the huggingface format and a custom plugin loads them.
    const res = await fetch(
      "https://api.mixpeek.com/v1/namespaces/ns_your_namespace/documents/upsert",
      {
        method: "POST",
        headers: {
          Authorization: "Bearer API_KEY",
          "Content-Type": "application/json",
        },
        body: JSON.stringify({
          collection_id: "col_your_collection",
          documents: [
            {
              document_id: "asset-00412",
              // The model produces text, so it lands in payload. Give the
              // collection a text vector index and embed that text to make it
              // searchable rather than only filterable.
              payload: { extracted_text: modelOutput, source_key: "archive/2026/asset-00412" },
              vectors: { "text-embedding": embeddingOfModelOutput },
            },
          ],
        }),
      },
    );
    
    // Managed alternative, if this exact model is not the requirement:
    // universal_extractor@v1 runs google/gemini-embedding-2
    // (3072-d) over a bucket, with no inference of your own.

    Capabilities

    • Self-supervised pre-training on unlabeled audio
    • Strong English ASR performance
    • Raw waveform input (no spectrogram needed)
    • Efficient fine-tuning with limited labeled data

    Use Cases on Mixpeek

    English-focused transcription workflows
    Low-resource language adaptation with limited training data
    Audio content indexing for search and discovery

    Benchmarks

    DatasetMetricScoreSource
    LibriSpeech (test-clean)WER2.7%Baevski et al., 2020: Table 5
    LibriSpeech (test-other)WER5.2%Baevski et al., 2020: Table 5

    Performance

    Input Sizevariable audio length
    GPU Latency~180ms / 30s chunk (A100)
    CPU Latency~2.8s / 30s chunk
    GPU Throughput~10× realtime (A100)
    GPU Memory~1.3 GB

    Specification

    FrameworkHF
    Organizationfacebook
    FeatureTranscription
    Outputtext + timestamps
    Modalitiesvideo, audio
    RetrieverTranscript Search
    Parameters317M
    Licenseapache-2.0
    Downloads/mo32K
    Likes35

    Research Paper

    wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations

    arxiv.org

    Build a pipeline with wav2vec2-large-960h

    Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

    Run it on your own data, free