Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
laibao
Qwen2.5_vllm
Commits
9919820f
Commit
9919820f
authored
Oct 11, 2024
by
laibao
Browse files
Update README.md
parent
54646388
Pipeline
#1754
failed with stages
in 0 seconds
Changes
1
Pipelines
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
6 additions
and
17 deletions
+6
-17
README.md
README.md
+6
-17
No files found.
README.md
View file @
9919820f
...
...
@@ -92,7 +92,7 @@ python examples/offline_inference.py
### 离线批量推理性能测试
1、指定输入输出
```
bash
python benchmarks/benchmark_throughput.py
--num-prompts
1
--input-len
32
--output-len
128
--model
Qwen/Qwen
1
.5-7B-Chat
-tp
1
--trust-remote-code
--enforce-eager
--dtype
float16
python benchmarks/benchmark_throughput.py
--num-prompts
1
--input-len
32
--output-len
128
--model
Qwen/Qwen
2
.5-7B-Chat
-tp
1
--trust-remote-code
--enforce-eager
--dtype
float16
```
其中
`--num-prompts`
是batch数,
`--input-len`
是输入seqlen,
`--output-len`
是输出token长度,
`--model`
为模型路径,
`-tp`
为使用卡数,
`dtype="float16"`
为推理数据类型,如果模型权重是bfloat16,需要修改为float16推理。若指定
`--output-len 1`
即为首字延迟。
`-q gptq`
为使用gptq量化模型进行推理。
...
...
@@ -103,28 +103,17 @@ wget https://huggingface.co/datasets/anon8231489123/ShareGPT_Vicuna_unfiltered/r
```
```
bash
python benchmarks/benchmark_throughput.py
--num-prompts
1
--model
Qwen/Qwen
1
.5-7B-Chat
--dataset
ShareGPT_V3_unfiltered_cleaned_split.json
-tp
1
--trust-remote-code
--enforce-eager
--dtype
float16
python benchmarks/benchmark_throughput.py
--num-prompts
1
--model
Qwen/Qwen
2
.5-7B-Chat
--dataset
ShareGPT_V3_unfiltered_cleaned_split.json
-tp
1
--trust-remote-code
--enforce-eager
--dtype
float16
```
其中
`--num-prompts`
是batch数,
`--model`
为模型路径,
`--dataset`
为使用的数据集,
`-tp`
为使用卡数,
`dtype="float16"`
为推理数据类型,如果模型权重是bfloat16,需要修改为float16推理。
`-q gptq`
为使用gptq量化模型进行推理。
### api服务推理性能测试
1、启动服务端:
```
bash
python
-m
vllm.entrypoints.openai.api_server
--model
Qwen/Qwen1.5-7B-Chat
--dtype
float16
--enforce-eager
-tp
1
```
2、启动客户端:
```
bash
python benchmarks/benchmark_serving.py
--model
Qwen/Qwen1.5-7B-Chat
--dataset
ShareGPT_V3_unfiltered_cleaned_split.json
--num-prompts
1
--trust-remote-code
```
参数同使用数据集,离线批量推理性能测试,具体参考
[
benchmarks/benchmark_serving.py
](
benchmarks/benchmark_serving.py
)
### OpenAI兼容服务
启动服务:
```
bash
python
-m
vllm.entrypoints.openai.api_server
--model
Qwen/Qwen
1
.5-7B-Chat
--enforce-eager
--dtype
float16
--trust-remote-code
python
-m
vllm.entrypoints.openai.api_server
--model
Qwen/Qwen
2
.5-7B-Chat
--enforce-eager
--dtype
float16
--trust-remote-code
```
这里
`--model`
为加载模型路径,
`--dtype`
为数据类型:float16,默认情况使用tokenizer中的预定义聊天模板,
`--chat-template`
可以添加新模板覆盖默认模板,
`-q gptq`
为使用gptq量化模型进行推理,
`-q awqq`
为使用awq量化模型进行推理。
...
...
@@ -138,7 +127,7 @@ curl http://localhost:8000/v1/models
curl http://localhost:8000/v1/completions
\
-H
"Content-Type: application/json"
\
-d
'{
"model": "Qwen/Qwen
1
.5-7B",
"model": "Qwen/Qwen
2
.5-7B",
"prompt": "What is deep learning?",
"max_tokens": 7,
"temperature": 0
...
...
@@ -152,7 +141,7 @@ curl http://localhost:8000/v1/completions \
curl http://localhost:8000/v1/chat/completions
\
-H
"Content-Type: application/json"
\
-d
'{
"model": "Qwen/Qwen
1
.5-7B-Chat",
"model": "Qwen/Qwen
2
.5-7B-Chat",
"messages": [
{"role": "system", "content": "What is deep learning?"},
{"role": "user", "content": "What is deep learning?"}
...
...
@@ -180,7 +169,7 @@ Prompt: 'What is deep learning?', Generated text: ' Deep learning is a subset of
金融,科研,教育
## 源码仓库及问题反馈
*
[
https://developer.hpccube.com/codes/modelzoo/qwen
1
.5_vllm
](
https://developer.hpccube.com/codes/modelzoo/qwen1.5_vllm
)
*
[
https://developer.hpccube.com/codes/modelzoo/qwen
2
.5_vllm
](
https://developer.hpccube.com/codes/modelzoo/qwen1.5_vllm
)
## 参考资料
*
[
https://github.com/vllm-project/vllm
](
https://github.com/vllm-project/vllm
)
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment