model_launch.sh

CUDA_VISIBLE_DEVICES=1 vllm serve Elfsong/VLM_stage_2_iter_0003500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.25 \
    --quantization bitsandbytes \
    --max_model_len 4096 \
    --port 9000 \
    --served-model-name Local-Model-03500

CUDA_VISIBLE_DEVICES=1 vllm serve Elfsong/VLM_stage_2_iter_0004500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.25 \
    --quantization bitsandbytes \
    --max_model_len 4096 \
    --port 9001 \
    --served-model-name Local-Model-04500

CUDA_VISIBLE_DEVICES=1 vllm serve Elfsong/VLM_stage_2_iter_0005500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.25 \
    --quantization bitsandbytes \
    --max_model_len 4096 \
    --port 9002 \
    --served-model-name Local-Model-05500

CUDA_VISIBLE_DEVICES=3 vllm serve Elfsong/VLM_stage_2_iter_0006500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.25 \
    --quantization bitsandbytes \
    --max_model_len 4096 \
    --port 9003 \
    --served-model-name Local-Model-06500

CUDA_VISIBLE_DEVICES=3 vllm serve Elfsong/VLM_stage_2_iter_0007500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.25 \
    --quantization bitsandbytes \
    --max_model_len 4096 \
    --port 9004 \
    --served-model-name Local-Model-07500

CUDA_VISIBLE_DEVICES=1 vllm serve Elfsong/VLM_stage_3_iter_0001500 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.3 \
    --quantization bitsandbytes \
    --seed 42 \
    --max_model_len 4096 \
    --port 9002 \
    --served-model-name Local-Model-01500

CUDA_VISIBLE_DEVICES=2 vllm serve Elfsong/VLM_stage_3_iter_0002000 \
    --tensor-parallel-size 1 \
    --gpu-memory-utilization 0.3 \
    --quantization bitsandbytes \
    --seed 42 \
    --max_model_len 4096 \
    --port 9003 \
    --served-model-name Local-Model-02000