DouDou commited on
Commit
a8b22df
·
verified ·
1 Parent(s): 115f0b5

Upload data3/run.sh with huggingface_hub

Browse files
Files changed (1) hide show
  1. data3/run.sh +12 -0
data3/run.sh ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ CUDA_VISIBLE_DEVICES=0,1
2
+
3
+ # vllm serve Qwen/Qwen3-0.6B-GPTQ-Int8 --tensor-parallel-size 2 --dtype auto --no-enable-chunked-prefill --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --quantization gptq --served-model-name Qwen3
4
+
5
+
6
+ # vllm serve Qwen/Qwen3-30B-A3B-GPTQ-Int4 --tensor-parallel-size 2 --dtype auto --no-enable-chunked-prefill --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --quantization gptq --served-model-name Qwen3
7
+
8
+ # vllm serve Qwen/Qwen3-14B-MLX-8bit --tensor-parallel-size 2 --dtype auto --no-enable-chunked-prefill --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --served-model-name Qwen3
9
+
10
+ # vllm serve Qwen/Qwen3-1.7B-GPTQ-Int8 --tensor-parallel-size 2 --dtype auto --no-enable-chunked-prefill --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --quantization gptq --served-model-name Qwen3
11
+
12
+ vllm serve Qwen/Qwen3-32B-AWQ --tensor-parallel-size 2 --dtype auto --no-enable-chunked-prefill --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}' --quantization awq --served-model-name Qwen3