config.json

{
  "model_type": "moss_ttsd",
  "architectures": [
    "MossTTSDForCausalLM"
  ],
  "auto_map": {
    "AutoProcessor": "processing_moss_ttsd.MossTTSDProcessor",
    "AutoConfig": "configuration_moss_ttsd.MossTTSDConfig",
    "AutoModel": "modeling_moss_ttsd.MossTTSDForCausalLM"
  },
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "channels": 8,
  "eos_token_id": 151643,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 6144,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "num_attention_heads": 16,
  "num_hidden_layers": 28,
  "num_key_value_heads": 8,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000,
  "sliding_window": null,
  "speech_pad_token": 1024,
  "pad_token": [
    152694,
    1024,
    1024,
    1024,
    1024,
    1024,
    1024,
    1024
  ],
  "speech_token_range": [
    151665,
    152689
  ],
  "speech_vocab_size": 1025,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.53.2",
  "use_cache": true,
  "use_sliding_window": false,
  "vocab_size": 152697,
  "vocab_size_list": [
    152697,
    1025,
    1025,
    1025,
    1025,
    1025,
    1025,
    1025
  ]
}