{ "architectures": [ "AudioVisualFlamingoForConditionalGeneration" ], "audio_chunk_length": 120, "audio_hop_length": 60, "audio_sampling_rate": 16000, "default_im_end_token": "", "default_im_start_token": "", "default_image_token": "", "default_sound_token": "", "dtype": "bfloat16", "dynamic_s2": true, "encoder_text_token_ids": { "\n": [ 198 ] }, "fps": 0.0, "hidden_size": 3584, "ignore_index": -100, "image_aspect_ratio": "dynamic_s2", "image_encoder": { "_target_": "BasicImageEncoder" }, "interleaved_video_segment_duration": 30, "interleaved_vis_aud_in_video": true, "interpolate_mode": "linear", "llm_cfg": { "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 151643, "eos_token_id": 151645, "hidden_act": "silu", "hidden_size": 3584, "initializer_range": 0.02, "intermediate_size": 18944, "max_position_embeddings": 32768, "max_window_layers": 28, "model_type": "qwen2", "num_attention_heads": 28, "num_hidden_layers": 28, "num_key_value_heads": 4, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 15300000.0, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "use_cache": false, "use_sliding_window": false, "vocab_size": 151661 }, "load_audio_in_video": true, "max_tiles": 12, "media_token_ids": { "image": 151649, "sound": 151652, "video": 151650 }, "media_tokens": { "image": "", "sound": "", "video": "" }, "min_tiles": 1, "mm_bos_eos_tokens": { "image": [ "<|image_bos|>", "<|image_eos|>" ], "sound": [ "<|sound_bos|>", "<|sound_eos|>" ], "video": [ "<|video_bos|>", "<|video_eos|>" ] }, "mm_hidden_size": 3456, "mm_projector_cfg": { "architectures": [ "MultimodalProjector" ], "mm_projector_type": "mlp_downsample", "model_type": "v2l_projector" }, "mm_use_im_patch_token": false, "mm_use_im_start_end": false, "mm_vision_select_feature": "cls_patch", "mm_vision_select_layer": -2, "model_type": "audiovisualflamingo", "num_time_tokens": 0, "num_video_frames": 32, "s2": false, "s2_max_split_size": 448, "s2_resize_output_to_scale_idx": -1, "s2_scales": "448,896,1344", "sentinel_token": "", "sound_encoder": { "_target_": "BasicSoundEncoder", "embed_time": "True", "end_tokens": "None", "time_embed_type": "lang", "trope_dim": 128, "trope_theta": 50000 }, "sound_hidden_size": 1280, "sound_mm_projector_cfg": { "architectures": [ "SoundMultimodalProjector" ], "model_type": "sound_mm_projector", "sound_mm_projector_type": "mlp" }, "sound_tower_cfg": { "activation_dropout": 0.0, "activation_function": "gelu", "architectures": [ "Qwen2AudioEncoder" ], "attention_dropout": 0.0, "d_model": 1280, "dropout": 0.0, "encoder_attention_heads": 20, "encoder_ffn_dim": 5120, "encoder_layerdrop": 0.0, "encoder_layers": 32, "init_std": 0.02, "max_source_positions": 1500, "model_type": "qwen2_audio_encoder", "num_hidden_layers": 32, "num_mel_bins": 128, "scale_embedding": false }, "time_token_format": "", "transformers_version": "5.3.0.dev0", "video_encoder": { "_target_": "TSPVideoEncoder", "embed_time": "True", "end_tokens": "None", "pool_sizes": [ [ 8, 1, 1 ] ], "time_embed_type": "lang", "trope_dim": 128, "trope_theta": 50000 }, "vision_resolution": -1, "vision_tower_cfg": { "architectures": [ "SiglipVisionModel" ], "attention_dropout": 0.0, "hidden_act": "gelu_pytorch_tanh", "hidden_size": 1152, "image_size": 448, "intermediate_size": 4304, "layer_norm_eps": 1e-06, "model_type": "siglip_vision_model", "num_attention_heads": 16, "num_channels": 3, "num_hidden_layers": 27, "num_image_tokens": 256, "patch_size": 14, "projection_dim": 2048, "projector_hidden_act": "gelu_fast", "vision_use_head": false } }