add multimodal LLM BLIP-2 and InstructBLIP
This commit is contained in:
parent
cde4dfe569
commit
4f3d558f67
|
@ -32,4 +32,3 @@ CUDA_VISIBLE_DEVICES=0 python src/train_bash.py \
|
|||
--quantization_bit 8 \
|
||||
--image_path /home/LAB/fengzc/LLM/checkpoints/liuhaotian/LLaVA-Instruct-150K/images/coco/train2017 \
|
||||
--use_qformer
|
||||
|
||||
|
|
Loading…
Reference in New Issue