av_model_v3 / config.json
Cyril666's picture
Upload folder using huggingface_hub
6691920 verified
{
"architectures": [
"Videollama3Qwen3ForCausalLM"
],
"auto_map": {
"AutoConfig": "configuration_videollama3.Videollama3Qwen3Config",
"AutoModelForCausalLM": "modeling_videollama3_qwen3.Videollama3Qwen3ForCausalLM"
},
"attention_bias": false,
"attention_dropout": 0.0,
"audio_encoder": "Cyril666/whisper-large-v3-encoder",
"audio_encoder_lr": 5e-06,
"audio_hidden_size": 1280,
"audio_projector_lr": 0.001,
"audio_projector_type": "dmlp2x_gelu",
"audio_token_index": 151672,
"bos_token_id": 151643,
"embedding_lr": null,
"eos_token_id": 151645,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 2048,
"image_aspect_ratio": "square",
"image_size": -1,
"image_token_index": 151669,
"image_token_length": 1,
"initializer_range": 0.02,
"intermediate_size": 6144,
"is_alignment": false,
"llm_lr": null,
"loss_reduction_scope": "batch",
"max_frames": 180,
"max_position_embeddings": 40960,
"max_window_layers": 28,
"mm_hidden_size": 1024,
"mm_projector_lr": 1e-05,
"mm_projector_type": "mlp2x_gelu",
"mm_vision_select_feature": "patch",
"mm_vision_select_layer": -1,
"model_type": "videollama3_qwen3",
"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,
"tie_word_embeddings": true,
"tokenizer_model_max_length": 16384,
"tokenizer_padding_side": "right",
"torch_dtype": "bfloat16",
"transformers_version": "4.51.3",
"use_cache": true,
"use_mm_proj": true,
"use_reconstruct": false,
"use_sliding_window": false,
"use_token_compression": false,
"use_vision_teacher": false,
"use_visual_expert": false,
"vision_encoder": null,
"vision_encoder_lr": null,
"vision_encoder_teacher": null,
"vision_hidden_size": 1024,
"vision_projector_lr": null,
"vision_projector_type": "mlp2x_gelu",
"visual_expert_lr": null,
"vocab_size": 151936,
"vision_encoder_config": {
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 1024,
"initializer_range": 0.02,
"intermediate_size": 3072,
"layer_norm_eps": 1e-06,
"max_window_layers": 28,
"num_attention_heads": 16,
"num_channels": 3,
"num_hidden_layers": 28,
"num_key_value_heads": 8,
"patch_size": 14,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 1000000,
"sliding_window": null,
"torch_dtype": "bfloat16"
}
}