config.json

{
    "model_name": "AutoModel",
    "hidden_size": 768,
    "num_attention_heads": 12,
    "num_hidden_layers": 12,
    "intermediate_size": 3072,
    "hidden_dropout_prob": 0.1,
    "attention_probs_dropout_prob": 0.1,
    "image_size": 224,
    "image_channels": 3,
    "patch_size": 16,
    "max_position_embeddings": 512,
    "vocab_size": 30522,
    "type_vocab_size": 2,
    "audio_sample_rate": 16000,
    "audio_frame_size": 1024,
    "audio_hop_size": 512,
    "enable_vqa": true,
    "enable_caption": true,
    "enable_retrieval": true,
    "enable_asr": true,
    "enable_realtime_asr": true,
    "batch_size": 32,
    "learning_rate": 0.0001,
    "weight_decay": 0.01,
    "warmup_steps": 10000,
    "max_steps": 100000
}