config.json

{
  "architectures": [
    "CM3PModel"
  ],
  "auto_map": {
    "AutoConfig": "configuration_cm3p.CM3PConfig",
    "AutoModel": "modeling_cm3p.CM3PModel"
  },
  "beatmap_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "audio_config": {
      "_name_or_path": "",
      "add_cross_attention": false,
      "architectures": null,
      "attention_bias": false,
      "attention_dropout": 0.0,
      "bad_words_ids": null,
      "begin_suppress_tokens": null,
      "bos_token_id": null,
      "chunk_size_feed_forward": 0,
      "cross_attention_hidden_size": null,
      "decoder_bias": true,
      "decoder_start_token_id": null,
      "deterministic_flash_attn": false,
      "diversity_penalty": 0.0,
      "do_sample": false,
      "early_stopping": false,
      "embedding_dropout": 0.0,
      "encoder_no_repeat_ngram_size": 0,
      "eos_token_id": null,
      "exponential_decay_length_penalty": null,
      "f_max": 8000,
      "f_min": 0,
      "finetuning_task": null,
      "forced_bos_token_id": null,
      "forced_eos_token_id": null,
      "global_attn_every_n_layers": 3,
      "global_rope_theta": 160000.0,
      "hidden_activation": "gelu",
      "hidden_size": 512,
      "hop_length": 128,
      "id2label": {
        "0": "LABEL_0",
        "1": "LABEL_1"
      },
      "initializer_cutoff_factor": 2.0,
      "initializer_range": 0.02,
      "intermediate_size": 1024,
      "is_decoder": false,
      "is_encoder_decoder": false,
      "label2id": {
        "LABEL_0": 0,
        "LABEL_1": 1
      },
      "length_penalty": 1.0,
      "local_attention": 128,
      "local_rope_theta": 10000.0,
      "max_length": 20,
      "max_position_embeddings": 4096,
      "min_length": 0,
      "mlp_bias": false,
      "mlp_dropout": 0.0,
      "model_type": "CM3PAudio",
      "n_ftt": 2048,
      "n_mels": 80,
      "no_repeat_ngram_size": 0,
      "norm_bias": false,
      "norm_eps": 1e-05,
      "num_attention_heads": 8,
      "num_beam_groups": 1,
      "num_beams": 1,
      "num_hidden_layers": 6,
      "num_return_sequences": 1,
      "output_attentions": false,
      "output_hidden_states": false,
      "output_scores": false,
      "pad_mode": "constant",
      "pad_token_id": null,
      "prefix": null,
      "problem_type": null,
      "projector_dim": 768,
      "projector_hidden_act": "gelu",
      "projector_intermediate_size": 2048,
      "pruned_heads": {},
      "remove_invalid_values": false,
      "repetition_penalty": 1.0,
      "return_dict": true,
      "return_dict_in_generate": false,
      "sample_rate": 16000,
      "sep_token_id": null,
      "suppress_tokens": null,
      "task_specific_params": null,
      "temperature": 1.0,
      "tf_legacy_loss": false,
      "tie_encoder_decoder": false,
      "tie_word_embeddings": true,
      "tokenizer_class": null,
      "top_k": 50,
      "top_p": 1.0,
      "torch_dtype": null,
      "torchscript": false,
      "typical_p": 1.0,
      "use_bfloat16": false,
      "vocab_size": 1
    },
    "audio_eos_token_id": 3966,
    "audio_sos_token_id": null,
    "audio_token_id": 3967,
    "bos_token_id": 3958,
    "classifier_activation": "gelu",
    "classifier_bias": false,
    "cls_embed": true,
    "decoder_bias": true,
    "deterministic_flash_attn": false,
    "embedding_dropout": 0.0,
    "eos_token_id": 3959,
    "global_attn_every_n_layers": 3,
    "global_rope_theta": 160000.0,
    "hidden_activation": "gelu",
    "hidden_size": 768,
    "initializer_cutoff_factor": 2.0,
    "initializer_factor": 1.0,
    "initializer_range": 0.02,
    "intermediate_size": 1152,
    "local_attention": 128,
    "local_rope_theta": 10000.0,
    "max_position_embeddings": 8192,
    "mlp_bias": false,
    "mlp_dropout": 0.0,
    "model_type": "CM3PBeatmap",
    "norm_bias": false,
    "norm_eps": 1e-05,
    "num_attention_heads": 12,
    "num_hidden_layers": 22,
    "pad_token_id": 3962,
    "projection_dim": 512,
    "repad_logits_with_grad": false,
    "sparse_pred_ignore_index": -100,
    "sparse_prediction": false,
    "torch_dtype": "bfloat16",
    "vocab_size": 3968
  },
  "has_decoder_head": true,
  "initializer_factor": 1.0,
  "initializer_range": 0.02,
  "logit_scale_init_value": 2.6592,
  "loss_type": "ForMaskedLM",
  "metadata_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 23127,
    "cls_embed": true,
    "decoder_bias": true,
    "deterministic_flash_attn": false,
    "embedding_dropout": 0.0,
    "eos_token_id": 23128,
    "global_attn_every_n_layers": 1,
    "global_rope_theta": 10000.0,
    "hidden_activation": "gelu",
    "hidden_size": 256,
    "initializer_cutoff_factor": 2.0,
    "initializer_factor": 1.0,
    "initializer_range": 0.02,
    "intermediate_size": 512,
    "local_attention": 128,
    "local_rope_theta": 10000.0,
    "max_position_embeddings": 128,
    "mlp_bias": false,
    "mlp_dropout": 0.0,
    "model_type": "CM3PMetadata",
    "norm_bias": false,
    "norm_eps": 1e-05,
    "num_attention_heads": 4,
    "num_hidden_layers": 6,
    "pad_token_id": 23129,
    "projection_dim": 512,
    "torch_dtype": "bfloat16",
    "vocab_size": 23145
  },
  "model_type": "CM3P",
  "projection_dim": 512,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.55.0"
}