When I try to use with --load-4bit, the description is completely incorrect. Do I need different weights?
python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40000 --worker http://localhost:40000 --model-path ./llava-hr-13b-x-sft-1024 --load-4bit
When I try to use with --load-4bit, the description is completely incorrect. Do I need different weights?
python -m llava.serve.model_worker --host 0.0.0.0 --controller http://localhost:10000 --port 40000 --worker http://localhost:40000 --model-path ./llava-hr-13b-x-sft-1024 --load-4bit