{
  "architectures": [
    "UpstreamFinetune"
  ],
  "classifier_output_dim": 9,
  "dropout": 0.2,
  "finetune_layers": 3,
  "hidden_dim": 64,
  "model_type": "wav2vec2-emodualhead",
  "num_layers": 2,
  "origin_upstream_url": "facebook/wav2vec2-base-960h",
  "regressor_output_dim": 2,
  "torch_dtype": "float32",
  "transformers_version": "4.52.3",
  "upstream_model": "wav2vec2-base-960h"
}