deepgeek/stacks
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(*):增加了一些模型

Browse Source
This commit is contained in:
deepgeek
2025-04-18 12:38:06 +08:00
parent a513dd8a91
commit d4c43315fb
19 changed files with 398 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
vllm-qwen25-72b-instruct-awq/compose.yaml Normal file → Executable file
View File

@@ -18,12 +18,12 @@ services:
runtime: nvidia
volumes:
- /home/deepgeek/data/data_local/server/vllm/models:/models
restart: always
command: --served-model-name Qwen2.5-72B-Instruct-AWQ --model /models/Qwen/Qwen2.5-72B-Instruct-AWQ
--trust-remote-code --host 0.0.0.0 --port 8080 --max-model-len 8192
--tensor-parallel-size 2 --gpu_memory_utilization 0.9 --enforce-eager
--dtype auto --swap-space 8 --enable-auto-tool-choice --tool-call-parser
llama3_json
restart: unless-stopped
command: --served-model-name Qwen2.5:72b --model
/models/Qwen/Qwen2.5-72B-Instruct-AWQ --trust-remote-code --host 0.0.0.0
--port 8080 --max-model-len 8192 --tensor-parallel-size 2
--gpu_memory_utilization 0.9 --enforce-eager --dtype auto --swap-space 8
--enable-auto-tool-choice --tool-call-parser llama3_json
x-dockge:
urls:
- http://local.citory.tech:10580