{
  "architectures": [
    "SheetSage2Model"
  ],
  "backbone_config": {
    "architectures": [
      "MERT2Model"
    ],
    "context_seconds": 360.0,
    "conv_depthwise_kernel_size": 31,
    "frame_rate": 25.0,
    "hidden_size": 1024,
    "hop_length": 240,
    "initializer_range": 0.02,
    "inputs_to_logits_ratio": 960,
    "intermediate_size": 4096,
    "layer_norm_eps": 1e-05,
    "minimum_input_samples": 1025,
    "model_type": "mert2",
    "n_fft": 2048,
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "num_mel_bins": 128,
    "rotary_embedding_base": 10000,
    "sampling_rate": 24000,
    "subsampling_channels": [
      128,
      512,
      1024
    ],
    "subsampling_depths": [
      3,
      4,
      5
    ],
    "subsampling_layer_norm_eps": 1e-06,
    "torch_dtype": "float32",
    "transformers_version": "4.53.2",
    "variant": "fs",
    "win_length": 2048
  },
  "base_model_name_or_path": "m-a-p/MERT-v2-FullSong",
  "base_model_revision": "d8ba1c745e733b3908ce6ad16ebeb17ac7600a42",
  "base_model_sha256": "e6dd2ab187d6dd62b6521cd7d8f932e237acf0c5757745a7232082e28391350d",
  "bos_token_id": 1,
  "decoder_dropout": 0.1,
  "decoder_layers": 6,
  "decoder_start_token_id": 1,
  "encoder_attn_implementation": "sdpa",
  "eos_token_id": 2,
  "hidden_size": 512,
  "input_audio_length": 300.0,
  "intermediate_size": 2048,
  "is_encoder_decoder": true,
  "lora_alpha": 128.0,
  "lora_rank": 64,
  "max_output_seq_len": 5120,
  "model_type": "sheetsage2",
  "num_attention_heads": 8,
  "pad_token_id": 0,
  "sampling_rate": 24000,
  "time_hz": 100,
  "tokenizer_fingerprint": "5ba3325af0344c7f",
  "tokenizer_schema_version": "v1",
  "transformers_version": "4.45.2",
  "use_cache": true,
  "vocab_size": 31678,
  "weights_format": "merged"
}