KORE: Enhancing Knowledge Injection for Large Multimodal Models via Knowledge-Oriented Augmentations and Constraints
Paper
•
2510.19316
•
Published
•
11
KORE
KORE uses knowledge-oriented control as its pivot to synergistically optimize the balance between knowledge adaptation and retention at different stag
-
-
kailinjiang/KORE-74K
Updated • 124 -
kailinjiang/llava_7b_ov_rank235
7B • Updated • 2 • 1 -
kailinjiang/llava_7b_ov_rank128
7B • Updated • 2 -
kailinjiang/llava_7b_ov_rank64
7B • Updated • 2 -
kailinjiang/llava_7b_ov_rank256
7B • Updated • 3 -
kailinjiang/llava_7b_ov_without_augmentation
7B • Updated • 2 • 1 -
kailinjiang/llava_7b_ov_without_constraint
7B • Updated • 3 -
kailinjiang/llava_7b_ov_without_frozen_matrix_A
7B • Updated • 1 -
kailinjiang/llava_7b_SEFE
7B • Updated • 5 -
kailinjiang/llava_7b_OLoRA
Updated • 2 -
kailinjiang/llava_13b_lora_rank128
13B • Updated • 1 -
kailinjiang/llava_13b_replay_lora_rank128
13B • Updated • 1 -
kailinjiang/Qwen2.5_VL_ov_rank274
8B • Updated • 2 -
kailinjiang/Qwen2.5_VL_7b_replay_lora
8B • Updated • 5 -
kailinjiang/Qwen2.5_VL_7b_lora
8B • Updated • 6 -
kailinjiang/SK_HallusionBench_rank235_llava
Updated • 9 -
kailinjiang/KO_Specific_Knowledge_MathVista_rank235_v2_llava_7b
Updated • 1 -
kailinjiang/KO_Specific_Knowledge_MME_rank235_v2_llava_7b
Updated • 2 -
kailinjiang/KO_Specific_Knowledge_OCRVQA_rank235_v2_llava_7b
7B • Updated • 1 -
kailinjiang/llava_13b_ov_rank235_h100
Updated • 5 -
kailinjiang/llava_7B_lora_image_knowledge_agnostic
7B • Updated • 2 -
kailinjiang/llava_7b_ov_training_base_ckpt_rank235
Updated • 14 -
kailinjiang/llava_7B_ewc
Updated • 5 -
kailinjiang/llava_7B_full_ft
Updated • 6 -
kailinjiang/llava_7B_lora
Updated • 2 -
kailinjiang/llava_7B_replay
Updated • 6