MonolithFoundation/Bumblebee
424
1{2 "_name_or_path": "./checkpoints/qwen-14b-chat",3 "adapt_size": false,4 "architectures": [5 "MiniGeminiQwen2ForCausalLM"6 ],7 "attention_dropout": 0.0,8 "bos_token_id": 151643,9 "eos_token_id": 151645,10 "freeze_mm_mlp_adapter": false,11 "hidden_act": "silu",12 "hidden_size": 5120,13 "image_aspect_ratio": "pad",14 "image_global": false,15 "image_grid": 1,16 "image_grid_pinpoints": null,17 "image_size_aux": 768,18 "initializer_range": 0.02,19 "intermediate_size": 13696,20 "max_position_embeddings": 32768,21 "max_window_layers": 35,22 "mm_hidden_size": 1024,23 "mm_hidden_size_aux": 2880,24 "mm_projector_lr": null,25 "mm_projector_type": "mlp2x_gelu",26 "mm_use_im_patch_token": false,27 "mm_use_im_start_end": false,28 "mm_vision_select_feature": "patch",29 "mm_vision_select_layer": -2,30 "mm_vision_tower": "./checkpoints/clip-vit-large-patch14-336",31 "mm_vision_tower_aux": "./checkpoints/openclip-convnext-large-d-320-laion2B-s29B-b131K-ft-soup",32 "model_type": "mini_gemini_qwen2",33 "new_img_size": null,34 "num_attention_heads": 40,35 "num_hidden_layers": 40,36 "num_key_value_heads": 40,37 "optimize_vision_tower": false,38 "optimize_vision_tower_aux": false,39 "rms_norm_eps": 1e-06,40 "rope_theta": 1000000.0,41 "sliding_window": 32768,42 "tie_word_embeddings": false,43 "tokenizer_model_max_length": 2048,44 "tokenizer_padding_side": "right",45 "torch_dtype": "bfloat16",46 "transformers_version": "4.39.3",47 "tune_mm_mlp_adapter": false,48 "use_cache": true,49 "use_mm_proj": true,50 "use_sliding_window": false,51 "vocab_size": 15206452}53 