{
  "architectures": [
    "MusicgenMelodyForConditionalGeneration"
  ],
  "audio_encoder": {
    "_name_or_path": "",
    "architectures": null,
    "audio_channels": 1,
    "chunk_length_s": null,
    "chunk_size_feed_forward": 0,
    "codebook_dim": 99,
    "codebook_size": 128,
    "compress": 1,
    "dilation_growth_rate": 2,
    "dtype": null,
    "hidden_size": 99,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "is_encoder_decoder": false,
    "kernel_size": 7,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "last_kernel_size": 7,
    "model_type": "encodec",
    "norm_type": "weight_norm",
    "normalize": false,
    "num_filters": 4,
    "num_lstm_layers": 2,
    "num_residual_layers": 1,
    "output_attentions": false,
    "output_hidden_states": false,
    "overlap": null,
    "pad_mode": "reflect",
    "problem_type": null,
    "residual_kernel_size": 3,
    "return_dict": true,
    "sampling_rate": 24000,
    "target_bandwidths": [
      1.5,
      3.0,
      6.0,
      12.0,
      24.0
    ],
    "trim_right_ratio": 1.0,
    "upsampling_ratios": [
      8,
      5,
      4,
      2
    ],
    "use_causal_conv": true,
    "use_conv_shortcut": true
  },
  "chroma_length": 24,
  "decoder": {
    "_name_or_path": "",
    "activation_dropout": 0.0,
    "activation_function": "gelu",
    "add_cross_attention": false,
    "architectures": null,
    "attention_dropout": 0.0,
    "audio_channels": 1,
    "bos_token_id": 99,
    "chunk_size_feed_forward": 0,
    "decoder_start_token_id": 99,
    "dropout": 0.1,
    "dtype": null,
    "eos_token_id": null,
    "ffn_dim": 4,
    "hidden_size": 16,
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "initializer_factor": 0.02,
    "is_decoder": false,
    "is_encoder_decoder": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layerdrop": 0.0,
    "max_position_embeddings": 2048,
    "model_type": "musicgen_melody_decoder",
    "num_attention_heads": 4,
    "num_codebooks": 4,
    "num_hidden_layers": 2,
    "output_attentions": false,
    "output_hidden_states": false,
    "pad_token_id": 99,
    "problem_type": null,
    "return_dict": true,
    "scale_embedding": false,
    "tie_word_embeddings": false,
    "use_cache": true,
    "vocab_size": 99
  },
  "dtype": "float32",
  "initializer_factor": 0.02,
  "is_encoder_decoder": true,
  "model_type": "musicgen_melody",
  "num_chroma": 12,
  "text_encoder": {
    "_name_or_path": "",
    "architectures": null,
    "chunk_size_feed_forward": 0,
    "classifier_dropout": 0.0,
    "d_ff": 4,
    "d_kv": 64,
    "d_model": 16,
    "dense_act_fn": "relu",
    "dropout_rate": 0.1,
    "dtype": null,
    "eos_token_id": 1,
    "feed_forward_proj": "relu",
    "id2label": {
      "0": "LABEL_0",
      "1": "LABEL_1"
    },
    "initializer_factor": 1.0,
    "is_decoder": false,
    "is_encoder_decoder": false,
    "is_gated_act": false,
    "label2id": {
      "LABEL_0": 0,
      "LABEL_1": 1
    },
    "layer_norm_epsilon": 1e-06,
    "model_type": "t5",
    "num_decoder_layers": 2,
    "num_heads": 4,
    "num_layers": 2,
    "output_attentions": false,
    "output_hidden_states": false,
    "pad_token_id": 0,
    "problem_type": null,
    "relative_attention_max_distance": 128,
    "relative_attention_num_buckets": 32,
    "return_dict": true,
    "scale_decoder_outputs": true,
    "tie_word_embeddings": true,
    "use_cache": false,
    "vocab_size": 99
  },
  "transformers_version": "5.16.0.dev0"
}