config.json

{
  "_name_or_path": "Lyra_Mini_3B",
  "align_norm": false,
  "align_temperature": 0.06,
  "align_type": "dtw",
  "architectures": [
    "Lyra2SQwen2VLForCausalLM"
  ],
  "asr_align": false,
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "check_data_modality": false,
  "compress": false,
  "compress_gap": 33,
  "ctc_decoder_config": "(2,4096,32,11008)",
  "ctc_loss_weight": 1.0,
  "ctc_upsample_factor": 25,
  "down_factor_1": 1,
  "down_factor_2": 1,
  "eos_token_id": 151645,
  "freeze_mm_mlp_adapter": true,
  "freeze_mm_speech_mlp_adapter": true,
  "hidden_act": "silu",
  "hidden_size": 1536,
  "image_aspect_ratio": "qwen2vl",
  "image_global": false,
  "image_grid": 1,
  "image_grid_pinpoints": null,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 8960,
  "keep_global": false,
  "keep_rate": 1.0,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "mm_hidden_size": 1536,
  "mm_projector_lr": null,
  "mm_projector_type": "identity",
  "mm_speech_hidden_size": 1280,
  "mm_speech_projector_type": "simple_mlp",
  "mm_speech_tower": "model_zoo/audio/whisper-large-v3-turbo",
  "mm_use_im_patch_token": false,
  "mm_use_im_start_end": false,
  "mm_use_mrope": false,
  "mm_vision_select_feature": "patch",
  "mm_vision_select_layer": -1,
  "mm_vision_tower": "model_zoo/vision/Qwen2VL_2B_ViT",
  "model_type": "Lyra_Qwen2VL_SpeechGenerator",
  "num_attention_heads": 12,
  "num_hidden_layers": 28,
  "num_key_value_heads": 2,
  "overlap_ratio": 0.0,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "speech_encoder_ds_rate": 5,
  "speech_encoder_hidden_size": 1280,
  "speech_generator_type": "ctc",
  "speech_learn": false,
  "speech_norm": false,
  "tie_word_embeddings": true,
  "tokenizer_model_max_length": 16384,
  "tokenizer_padding_side": "right",
  "torch_dtype": "bfloat16",
  "train_modality": "text_speech",
  "transformers_version": "4.45.0",
  "tune_mm_mlp_adapter": false,
  "tune_mm_speech_mlp_adapter": false,
  "tune_speech_generator_only": true,
  "unit_vocab_size": 1000,
  "use_cache": true,
  "use_mm_proj": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "hidden_size": 1536,
    "in_chans": 3,
    "model_type": "qwen2_vl",
    "spatial_patch_size": 14
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 151936,
  "weight_lambda": 0.0
}