jedyang97's picture
correct model loading logic
f2c4ee8
raw
history blame
1.62 kB
{
"_name_or_path": "/nfs/turbo/coe-chaijy-unreplicated/pre-trained-weights/LLaVA/llava-llama-2-7b-chat-lightning-preview",
"added_special_token_to_input_id": {},
"architectures": [
"LlavaLlamaForCausalLM"
],
"bbox_ce_loss_weight": 0.0,
"bbox_iou_loss_weight": 0.0,
"bbox_mse_loss_weight": 0.0,
"bbox_tokenization_type": null,
"bos_token_id": 1,
"d_latents": 1024,
"eos_token_id": 2,
"freeze_mm_mlp_adapter": false,
"hidden_act": "silu",
"hidden_size": 4096,
"image_aspect_ratio": "square",
"image_grid_pinpoints": null,
"initializer_range": 0.02,
"intermediate_size": 11008,
"lm_loss_weight": 1.0,
"max_position_embeddings": 2048,
"mm_hidden_size": 1024,
"mm_resampler_type": null,
"mm_use_im_patch_token": false,
"mm_use_im_start_end": false,
"mm_vision_select_feature": "patch",
"mm_vision_select_layer": -2,
"mm_vision_tower": "bbox-ground-truth",
"model_type": "llava",
"normalize_points_to_unit_cube": false,
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 32,
"num_latents": 256,
"num_voxels_per_axis_for_location_tokens": 10,
"pad_token_id": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.31.0",
"tune_mm_mlp_adapter": false,
"tune_mm_vision_resampler": false,
"tune_vision_tower": false,
"use_bbox_ce_loss": false,
"use_bbox_iou_loss": false,
"use_bbox_mse_loss": false,
"use_cache": false,
"use_mm_proj": true,
"vision_tower": "bbox-ground-truth",
"vocab_size": 32000
}