{
  "activation_dropout": 0.0,
  "architectures": [
    "GraniteSpeech5Encoder"
  ],
  "attention_bias": true,
  "attention_dropout": 0.0,
  "context_size": 13,
  "conv_expansion_factor": 2,
  "conv_kernel_size": 7,
  "dtype": "float32",
  "head_dim": 16,
  "hidden_act": "silu",
  "hidden_size": 32,
  "initializer_range": 0.02,
  "intermediate_size": 64,
  "max_position_embeddings": 200,
  "model_type": "granite_speech5_encoder",
  "num_attention_heads": 2,
  "num_hidden_layers": 2,
  "num_key_value_heads": 2,
  "num_mel_bins": 6,
  "subsample_layers": [
    0
  ],
  "transformers_version": "5.19.0.dev0",
  "vocab_size": 16384
}