config.json

{
    "model_type": "speech_to_text_2",
    "architectures": ["Speech2Text2ForConditionalGeneration"],
    "activation_dropout": 0.1,
    "activation_function": "relu",
    "attention_dropout": 0.1,
    "d_model": 512,
    "decoder_attention_heads": 8,
    "decoder_ffn_dim": 2048,
    "decoder_layers": 6,
    "dropout": 0.1,
    "encoder_attention_heads": 8,
    "encoder_ffn_dim": 2048,
    "encoder_layers": 6,
    "init_std": 0.02,
    "max_speech_positions": 4000,
    "max_text_positions": 1024,
    "num_conv_layers": 2,
    "num_hidden_layers": 12,
    "speech_vocab_size": 4096,
    "vocab_size": 50265,
    "use_cache": true,
    "tie_word_embeddings": true,
    "is_encoder_decoder": true,
    "pad_token_id": 1,
    "bos_token_id": 0,
    "eos_token_id": 2,
    "transformers_version": "4.31.0"
}