EngineX-Iluvatar/mr_v100-vllm
15
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Compare commits

merge into: EngineX-Iluvatar:main
Branches Tags
EngineX-Iluvatar:main EngineX-Iluvatar:docs-readme
...
pull from: EngineX-Iluvatar:docs-readme
Branches Tags
EngineX-Iluvatar:main EngineX-Iluvatar:docs-readme

1 Commits
main ... docs-readm

Author SHA1 Message Date
aceforeverd
82001b8ff0 update model list 2025-12-03 15:14:10 +08:00
1 changed files with 27 additions and 25 deletions
Expand all files Collapse all files

52
README.md
View File

@@ -110,28 +110,30 @@ print(response.choices[0].message.content)
在相同模型和输入条件下，测试平均输出速度（单位：字每秒），结果如下：
| 模型 | 智铠100 输出速度 | Nvidia A100 输出速度 |
|--------|--------------------------|-------------------------------|
| Qwen2.5-7B-Instruct | 56.4 | 112.4 |
| Qwen2.5-1.5B-Instruct-AWQ | 123.1 | 100.8 |
| Qwen/Qwen2.5-3B-Instruct | 91.7 | 95.6 |
| Qwen/Qwen2-7B-Instruct | 59.2 | 110.1 |
| Qwen/Qwen2-7B | 74.3 | 169.9 |
| Qwen/Qwen2-1.5B | 161.2 | 175.5 |
| Qwen/Qwen2-0.5B-Instruct | 141.1 | 146.2 |
| Qwen/Qwen2-1.5B-Instruct | 119.4 | 124.5 |
| Qwen/Qwen1.5-4B-Chat | 87.0 | 95.5 |
| Qwen/Qwen1.5-14B-Chat | 52.1 | 72.2 |
| Qwen/Qwen-1_8B-Chat | 151.6 | 203.9 |
| Qwen/Qwen-7B | 90.4 | 112.5 |
| Qwen/Qwen-7B-Chat | 93.7 | 131.4 |
| X-D-Lab/MindChat-Qwen-7B | 92.9 | 123.4 |
| Qwen/Qwen2.5-32B-Instruct-AWQ | 52.6 | 47.4 |
| Qwen/CodeQwen1.5-7B-Chat | 98.9 | 108.1 |
| Qwen/Qwen2.5-72B-Instruct-AWQ | 30.7 | 41.4 |
| Valdemardi/DeepSeek-R1-Distill-Qwen-32B-AWQ | 51.9 | 51.6 |
| Qwen/Qwen3-32B-AWQ | 31.8 | 44.1 |
| Qwen/QwQ-32B-AWQ | 50.9 | 48.0 |
| swift/Qwen3-30B-A3B-AWQ | 29.5 | 38.1 |
| Qwen/Qwen3-14B-AWQ | 53.4 | 62.6 |
| codefuse-ai/CodeFuse-QWen-14B | 61.6 | 75.5 |
| 模型名称 | A100出字速度(字/秒) | 出字速度(字/秒) | A100输出质量 | 输出质量 | A100首字延迟(秒) | 首字延迟(秒) | 备注 |
| ----- | ----- | ----- | ----- | ----- | ----- | ----- | ----- |
| AI-ModelScope/Mistral-7B-Instruct-v0.2 | 50.0638 | 55.3545 | 88.5000 | 85.0000 | 0.2544 | 0.2078 | |
| codefuse-ai/CodeFuse-QWen-14B | 75.4755 | 61.5555 | 63.7500 | 42.5000 | 0.1983 | 0.0840 | |
| Qwen/CodeQwen1.5-7B-Chat | 108.0509 | 98.9290 | 35.0000 | 23.7500 | 0.1042 | 0.1735 | |
| Qwen/Qwen-1_8B-Chat | 203.9495 | 151.5506 | 65.0000 | 60.0000 | 0.0760 | 0.0951 | |
| Qwen/Qwen-7B | 112.5454 | 90.3540 | 55.0000 | 63.7500 | 0.1047 | 0.1419 | |
| Qwen/Qwen-7B-Chat | 131.4390 | 93.7376 | 85.0000 | 65.0000 | 0.1198 | 0.0920 | |
| Qwen/Qwen1.5-1.8B | 104.2267 | 111.8826 | 26.2500 | 37.5000 | 0.1176 | 0.0990 | |
| Qwen/Qwen1.5-14B-Chat | 72.2311 | 52.0573 | 89.7500 | 88.5000 | 0.1024 | 0.1990 | |
| Qwen/Qwen1.5-4B-Chat | 95.4927 | 87.0352 | 85.0000 | 85.0000 | 0.0991 | 0.1340 | |
| Qwen/Qwen2-0.5B-Instruct | 146.2331 | 141.0503 | 50.0000 | 56.2500 | 0.0981 | 0.1003 | |
| Qwen/Qwen2-1.5B | 175.4972 | 161.2327 | 38.7500 | 53.7500 | 0.1036 | 0.1298 | |
| Qwen/Qwen2-1.5B-Instruct | 124.5098 | 119.4177 | 75.0000 | 80.0000 | 0.0895 | 0.1117 | |
| Qwen/Qwen2-7B | 169.9027 | 74.3288 | 72.5000 | 56.2500 | 0.1120 | 0.2810 | |
| Qwen/Qwen2-7B-Instruct | 110.1237 | 59.1989 | 89.2500 | 89.7500 | 0.0971 | 0.2078 | |
| Qwen/Qwen2.5-1.5B-Instruct | 116.6704 | 123.0560 | 85.0000 | 86.7500 | 0.1291 | 0.1216 | |
| Qwen/Qwen2.5-32B-Instruct-AWQ | 47.4427 | 52.5942 | 91.0000 | 87.5000 | 0.1332 | 0.3550 | |
| Qwen/Qwen2.5-3B-Instruct | 95.6249 | 91.6548 | 85.0000 | 85.0000 | 0.1122 | 0.1332 | |
| Qwen/Qwen2.5-72B-Instruct-AWQ | 41.4106 | 30.7387 | 91.0000 | 91.0000 | 0.1366 | 0.7175 | |
| Qwen/Qwen2.5-VL-7B-Instruct-AWQ | 61.5433 | 56.6830 | 88.5000 | 88.5000 | 0.2346 | 0.2472 | |
| Qwen/Qwen3-14B-AWQ | 62.6335 | 53.3993 | 86.7500 | 88.5000 | 0.1429 | 0.2347 | |
| Qwen/Qwen3-32B-AWQ | 44.0649 | 31.8064 | 89.2500 | 88.5000 | 0.2027 | 0.5392 | |
| Qwen/QwQ-32B-AWQ | 47.9752 | 50.8763 | 88.5000 | 88.5000 | 0.2201 | 0.4284 | |
| swift/Qwen3-30B-A3B-AWQ | 38.1391 | 29.5163 | 88.0000 | 88.5000 | 0.1619 | 0.2017 | |
| Valdemardi/DeepSeek-R1-Distill-Qwen-32B-AWQ | 51.5753 | 51.8930 | 88.0000 | 88.0000 | 0.2052 | 0.4199 | |
| X-D-Lab/MindChat-Qwen-7B | 123.3664 | 92.9101 | 71.2500 | 70.0000 | 0.1032 | 0.1271 | |