{
  "format": "positiveloss-weights",
  "architecture": "w2v-bert",
  "source": "/home/yehor/Work/PositiveLoss/squeezeformer-ukrainian/w2v/model.safetensors",
  "weights": "/home/yehor/Work/PositiveLoss/squeezeformer-ukrainian/w2v-weights/model.safetensors",
  "tensor_count": 797,
  "tensor_shapes": {
    "lm_head.bias": [
      512
    ],
    "lm_head.weight": [
      512,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.ffn.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.adapter.layers.0.ffn.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.ffn.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.ffn.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.adapter.layers.0.ffn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.ffn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.residual_conv.bias": [
      2048
    ],
    "wav2vec2_bert.adapter.layers.0.residual_conv.weight": [
      2048,
      1024,
      3
    ],
    "wav2vec2_bert.adapter.layers.0.residual_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.residual_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn_conv.bias": [
      2048
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn_conv.weight": [
      2048,
      1024,
      3
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.adapter.layers.0.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.0.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.0.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.1.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.1.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.10.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.10.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.11.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.11.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.12.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.12.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.13.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.13.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.14.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.14.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.15.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.15.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.16.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.16.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.17.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.17.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.18.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.18.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.19.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.19.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.2.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.2.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.20.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.20.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.21.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.21.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.22.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.22.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.23.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.23.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.3.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.3.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.4.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.4.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.5.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.5.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.6.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.6.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.7.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.7.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.8.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.8.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.depthwise_conv.weight": [
      1024,
      1,
      31
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.depthwise_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.depthwise_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.pointwise_conv1.weight": [
      2048,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.9.conv_module.pointwise_conv2.weight": [
      1024,
      1024,
      1
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn1_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2.intermediate_dense.bias": [
      4096
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2.intermediate_dense.weight": [
      4096,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2.output_dense.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2.output_dense.weight": [
      1024,
      4096
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.ffn2_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.final_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.final_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.distance_embedding.weight": [
      73,
      64
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_k.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_k.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_out.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_out.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_q.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_q.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_v.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn.linear_v.weight": [
      1024,
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn_layer_norm.bias": [
      1024
    ],
    "wav2vec2_bert.encoder.layers.9.self_attn_layer_norm.weight": [
      1024
    ],
    "wav2vec2_bert.feature_projection.layer_norm.bias": [
      160
    ],
    "wav2vec2_bert.feature_projection.layer_norm.weight": [
      160
    ],
    "wav2vec2_bert.feature_projection.projection.bias": [
      1024
    ],
    "wav2vec2_bert.feature_projection.projection.weight": [
      1024,
      160
    ],
    "wav2vec2_bert.masked_spec_embed": [
      1024
    ]
  },
  "encoder_config": null,
  "training_args": null
}