config.json

{
  "architectures": [
    "MiniCPMO"
  ],
  "attention_dropout": 0.0,
  "audio_chunk_length": 1.0,
  "audio_config": {
    "_name_or_path": "openai/whisper-medium",
    "activation_dropout": 0.0,
    "activation_function": "gelu",
    "apply_spec_augment": false,
    "architectures": [
      "MiniCPMWhisperEncoder"
    ],
    "attention_dropout": 0.0,
    "begin_suppress_tokens": [
      0
    ],
    "bos_token_id": 0,
    "classifier_proj_size": 256,
    "d_model": 8,
    "decoder_attention_heads": 1,
    "decoder_ffn_dim": 16,
    "decoder_layerdrop": 0.0,
    "decoder_layers": 1,
    "decoder_start_token_id": 50258,
    "dropout": 0.0,
    "dtype": "float32",
    "encoder_attention_heads": 1,
    "encoder_ffn_dim": 16,
    "encoder_layerdrop": 0.0,
    "encoder_layers": 1,
    "eos_token_id": 1,
    "forced_decoder_ids": [],
    "init_std": 0.02,
    "mask_feature_length": 10,
    "mask_feature_min_masks": 0,
    "mask_feature_prob": 0.0,
    "mask_time_length": 10,
    "mask_time_min_masks": 2,
    "mask_time_prob": 0.05,
    "max_length": 448,
    "max_source_positions": 1500,
    "max_target_positions": 448,
    "median_filter_width": 7,
    "model_type": "whisper",
    "num_hidden_layers": 24,
    "num_mel_bins": 80,
    "pad_token_id": 2,
    "scale_embedding": false,
    "suppress_tokens": [],
    "use_cache": true,
    "use_weighted_layer_sum": false,
    "vocab_size": 1000
  },
  "audio_pool_step": 2,
  "auto_map": {
    "AutoConfig": "configuration_minicpm.MiniCPMOConfig",
    "AutoModel": "modeling_minicpmo.MiniCPMO",
    "AutoModelForCausalLM": "modeling_minicpmo.MiniCPMO"
  },
  "batch_vision_input": true,
  "bos_token_id": 0,
  "chunk_input": true,
  "drop_vision_last_layer": false,
  "dtype": "bfloat16",
  "eos_token_id": 1,
  "hidden_act": "silu",
  "hidden_size": 16,
  "image_size": 448,
  "init_audio": true,
  "init_tts": false,
  "init_vision": true,
  "initializer_range": 0.02,
  "intermediate_size": 32,
  "layer_types": [
    "full_attention"
  ],
  "listen_speak_type": "asr",
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "minicpmo",
  "num_attention_heads": 4,
  "num_hidden_layers": 1,
  "num_key_value_heads": 1,
  "pad_token_id": 2,
  "patch_size": 14,
  "query_num": 64,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "slice_config": {
    "max_slice_nums": 9,
    "model_type": "minicpmv",
    "patch_size": 14,
    "scale_resolution": 448
  },
  "slice_mode": true,
  "sliding_window": null,
  "stream_input": false,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.6",
  "tts_config": {
    "attn_implementation": "sdpa",
    "audio_bos_token_id": 21132,
    "aug_loss_weight": true,
    "hidden_size": 768,
    "intermediate_size": 3072,
    "llm_dim": 3584,
    "max_position_embeddings": 4096,
    "model_type": "conditional_chattts",
    "num_attention_heads": 12,
    "num_audio_tokens": 626,
    "num_hidden_layers": 20,
    "num_mel_bins": 100,
    "num_spk_embs": 1,
    "num_text_tokens": 21178,
    "num_vq": 4,
    "spk_emb_token_id": 21143,
    "streaming": true,
    "streaming_audio_chunk_size": 50,
    "streaming_text_chunk_size": 10,
    "streaming_text_reserved_len": 300,
    "text_eos_token_id": 21133,
    "use_llm_hidden_state": false,
    "use_mlp": true,
    "use_speaker_embedding": true,
    "use_text": true
  },
  "use_cache": true,
  "use_image_id": true,
  "use_sliding_window": false,
  "version": 2.6,
  "vision_batch_size": 16,
  "vision_config": {
    "attention_dropout": 0.0,
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 8,
    "image_size": 980,
    "intermediate_size": 16,
    "layer_norm_eps": 1e-06,
    "model_type": "siglip_vision_model",
    "num_attention_heads": 1,
    "num_channels": 3,
    "num_hidden_layers": 1,
    "patch_size": 14
  },
  "vocab_size": 32000
}