dimavz/
whisper-tiny

Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation.

2,007 6 weeks ago

1 Tag
... /
model
d76121b83ea6 · 44MB
    Metadata
    Tensor
  • model.decoder.embed_positions.weight
    F32
    [384, 448]
  • model.decoder.embed_tokens.weight
    Q8_0
    [384, 51865]
  • model.decoder.layer_norm.bias
    F32
    [384]
  • model.decoder.layer_norm.weight
    F32
    [384]
  • model.decoder.layers.0.encoder_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.encoder_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.0.encoder_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.encoder_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.0.encoder_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.encoder_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.0.encoder_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.encoder_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.0.encoder_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.0.fc1.bias
    F32
    [1536]
  • model.decoder.layers.0.fc1.weight
    Q8_0
    [384, 1536]
  • model.decoder.layers.0.fc2.bias
    F32
    [384]
  • model.decoder.layers.0.fc2.weight
    Q8_0
    [1536, 384]
  • model.decoder.layers.0.final_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.0.final_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.0.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.self_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.0.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.self_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.0.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.self_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.0.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.0.self_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.0.self_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.1.encoder_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.encoder_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.1.encoder_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.encoder_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.1.encoder_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.encoder_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.1.encoder_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.encoder_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.1.encoder_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.1.fc1.bias
    F32
    [1536]
  • model.decoder.layers.1.fc1.weight
    Q8_0
    [384, 1536]
  • model.decoder.layers.1.fc2.bias
    F32
    [384]
  • model.decoder.layers.1.fc2.weight
    Q8_0
    [1536, 384]
  • model.decoder.layers.1.final_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.1.final_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.1.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.self_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.1.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.self_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.1.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.self_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.1.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.1.self_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.1.self_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.2.encoder_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.encoder_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.2.encoder_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.encoder_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.2.encoder_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.encoder_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.2.encoder_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.encoder_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.2.encoder_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.2.fc1.bias
    F32
    [1536]
  • model.decoder.layers.2.fc1.weight
    Q8_0
    [384, 1536]
  • model.decoder.layers.2.fc2.bias
    F32
    [384]
  • model.decoder.layers.2.fc2.weight
    Q8_0
    [1536, 384]
  • model.decoder.layers.2.final_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.2.final_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.2.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.self_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.2.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.self_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.2.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.self_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.2.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.2.self_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.2.self_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.3.encoder_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.encoder_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.3.encoder_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.encoder_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.3.encoder_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.encoder_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.3.encoder_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.encoder_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.3.encoder_attn_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.3.fc1.bias
    F32
    [1536]
  • model.decoder.layers.3.fc1.weight
    Q8_0
    [384, 1536]
  • model.decoder.layers.3.fc2.bias
    F32
    [384]
  • model.decoder.layers.3.fc2.weight
    Q8_0
    [1536, 384]
  • model.decoder.layers.3.final_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.3.final_layer_norm.weight
    F32
    [384]
  • model.decoder.layers.3.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.self_attn.out_proj.bias
    F32
    [384]
  • model.decoder.layers.3.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.self_attn.q_proj.bias
    F32
    [384]
  • model.decoder.layers.3.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.self_attn.v_proj.bias
    F32
    [384]
  • model.decoder.layers.3.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.decoder.layers.3.self_attn_layer_norm.bias
    F32
    [384]
  • model.decoder.layers.3.self_attn_layer_norm.weight
    F32
    [384]
  • model.encoder.conv1.bias
    F32
    [384]
  • model.encoder.conv1.weight
    F32
    [3, 80, 384]
  • model.encoder.conv2.bias
    F32
    [384]
  • model.encoder.conv2.weight
    F32
    [3, 384, 384]
  • model.encoder.embed_positions.weight
    F32
    [384, 1500]
  • model.encoder.layer_norm.bias
    F32
    [384]
  • model.encoder.layer_norm.weight
    F32
    [384]
  • model.encoder.layers.0.fc1.bias
    F32
    [1536]
  • model.encoder.layers.0.fc1.weight
    Q8_0
    [384, 1536]
  • model.encoder.layers.0.fc2.bias
    F32
    [384]
  • model.encoder.layers.0.fc2.weight
    Q8_0
    [1536, 384]
  • model.encoder.layers.0.final_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.0.final_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.0.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.0.self_attn.out_proj.bias
    F32
    [384]
  • model.encoder.layers.0.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.0.self_attn.q_proj.bias
    F32
    [384]
  • model.encoder.layers.0.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.0.self_attn.v_proj.bias
    F32
    [384]
  • model.encoder.layers.0.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.0.self_attn_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.0.self_attn_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.1.fc1.bias
    F32
    [1536]
  • model.encoder.layers.1.fc1.weight
    Q8_0
    [384, 1536]
  • model.encoder.layers.1.fc2.bias
    F32
    [384]
  • model.encoder.layers.1.fc2.weight
    Q8_0
    [1536, 384]
  • model.encoder.layers.1.final_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.1.final_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.1.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.1.self_attn.out_proj.bias
    F32
    [384]
  • model.encoder.layers.1.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.1.self_attn.q_proj.bias
    F32
    [384]
  • model.encoder.layers.1.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.1.self_attn.v_proj.bias
    F32
    [384]
  • model.encoder.layers.1.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.1.self_attn_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.1.self_attn_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.2.fc1.bias
    F32
    [1536]
  • model.encoder.layers.2.fc1.weight
    Q8_0
    [384, 1536]
  • model.encoder.layers.2.fc2.bias
    F32
    [384]
  • model.encoder.layers.2.fc2.weight
    Q8_0
    [1536, 384]
  • model.encoder.layers.2.final_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.2.final_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.2.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.2.self_attn.out_proj.bias
    F32
    [384]
  • model.encoder.layers.2.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.2.self_attn.q_proj.bias
    F32
    [384]
  • model.encoder.layers.2.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.2.self_attn.v_proj.bias
    F32
    [384]
  • model.encoder.layers.2.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.2.self_attn_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.2.self_attn_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.3.fc1.bias
    F32
    [1536]
  • model.encoder.layers.3.fc1.weight
    Q8_0
    [384, 1536]
  • model.encoder.layers.3.fc2.bias
    F32
    [384]
  • model.encoder.layers.3.fc2.weight
    Q8_0
    [1536, 384]
  • model.encoder.layers.3.final_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.3.final_layer_norm.weight
    F32
    [384]
  • model.encoder.layers.3.self_attn.k_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.3.self_attn.out_proj.bias
    F32
    [384]
  • model.encoder.layers.3.self_attn.out_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.3.self_attn.q_proj.bias
    F32
    [384]
  • model.encoder.layers.3.self_attn.q_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.3.self_attn.v_proj.bias
    F32
    [384]
  • model.encoder.layers.3.self_attn.v_proj.weight
    Q8_0
    [384, 384]
  • model.encoder.layers.3.self_attn_layer_norm.bias
    F32
    [384]
  • model.encoder.layers.3.self_attn_layer_norm.weight
    F32
    [384]