wehub-resource-sync
11 lines
400 B
Bash
11 lines
400 B
Bash
CUDA_VISIBLE_DEVICES=0,1,2,3 \
|
|
swift infer \
|
|
--model Qwen/Qwen3-Next-80B-A3B-Instruct \
|
|
--vllm_tensor_parallel_size 4 \
|
|
--infer_backend vllm \
|
|
--vllm_max_model_len 8192 \
|
|
--val_dataset AI-ModelScope/alpaca-gpt4-data-zh#100 \
|
|
--vllm_speculative_config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}' \
|
|
--vllm_gpu_memory_utilization 0.9 \
|
|
--max_new_tokens 2048
|