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
ModelZoo
Qwen2.5-VL_pytorch
Commits
c5d362f1
Commit
c5d362f1
authored
Jan 26, 2026
by
raojy
Browse files
updata readme
parent
408de0f5
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
46 additions
and
20 deletions
+46
-20
README.md
README.md
+46
-20
No files found.
README.md
View file @
c5d362f1
...
...
@@ -40,10 +40,41 @@ Qwen2.5-VL 从头开始训练了一个原生动态分辨率的 ViT,包括 CLIP
推荐使用镜像: harbor.sourcefind.cn:5443/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.2-1226-das1.7-py3.10-20251226
-
挂载地址
`-v`
根据实际模型情况修改
-
挂载地址
`-v`
,
`{docker_name}`
和
`{docker_image_name}`
根据实际模型情况修改
```
bash
docker run
-it
--shm-size
60g
--network
=
host
--name
minimax_m2
--privileged
--device
=
/dev/kfd
--device
=
/dev/dri
--device
=
/dev/mkfd
--group-add
video
--cap-add
=
SYS_PTRACE
--security-opt
seccomp
=
unconfined
-u
root
-v
/opt/hyhal/:/opt/hyhal/:ro
-v
/path/your_code_path/:/path/your_code_path/ harbor.sourcefind.cn:5443/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.2-1226-das1.7-py3.10-20251226
docker run
-it
\
--shm-size
60g
\
--network
=
host
\
--name
{
docker_name
}
\
--privileged
\
--device
=
/dev/kfd
\
--device
=
/dev/dri
\
--device
=
/dev/mkfd
\
--group-add
video
\
--cap-add
=
SYS_PTRACE
\
--security-opt
seccomp
=
unconfined
\
-u
root
\
-v
/opt/hyhal/:/opt/hyhal/:ro
\
-v
/path/your_code_data/:/path/your_code_data/
\
{
docker_image_name
}
bash
示例如下
(
展示到modelzoo上的内容,就是将上面的
{
docker_image_name
}{
docker_name
}
根据实际模型填写
)
:
docker run
-it
\
--shm-size
60g
\
--network
=
host
\
--name
qwen3
\
--privileged
\
--device
=
/dev/kfd
\
--device
=
/dev/dri
\
--device
=
/dev/mkfd
\
--group-add
video
\
--cap-add
=
SYS_PTRACE
\
--security-opt
seccomp
=
unconfined
\
-u
root
\
-v
/opt/hyhal/:/opt/hyhal/:ro
\
-v
/path/your_code_data/:/path/your_code_data/
\
image.sourcefind.cn:5000/dcu/admin/base/vllm:0.9.2-ubuntu22.04-dtk25.04.2-py3.10 bash
```
更多镜像可前往
[
光源
](
https://sourcefind.cn/#/service-list
)
下载使用。
...
...
@@ -200,6 +231,10 @@ torchrun ./LLaMA-Factory/src/train.py \
## 推理
1.
推理框架有
`transformers`
、
`vllm`
、其他推理框架中任意一个即可,至少有一个;
### transformers
### 单机单卡
```
...
...
@@ -233,10 +268,6 @@ CUDA_VISIBLE_DEVICES=0,1,2,3 python inference.py
#### 单机推理
```
## serve启动
export ALLREDUCE_STREAM_WITH_COMPUTE=1
export VLLM_MLA_DISABLE=0
export VLLM_USE_FLASH_MLA=1
# 启动命令
vllm serve Qwen/Qwen2.5-VL-3B-Instruct \
...
...
@@ -251,7 +282,7 @@ vllm serve Qwen/Qwen2.5-VL-3B-Instruct \
curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "
minimax
",
"model": "
qwen-vl
",
"messages": [
{
"role": "user",
...
...
@@ -272,23 +303,18 @@ curl http://localhost:8000/v1/chat/completions \
DCU与GPU精度一致,推理框架:vllm。
## 应用场景
### 算法类别
`对话问答`
### 热点应用行业
`科研,教育,政府,金融`
## 预训练权重
[
ModelScope
](
https://modelscope.cn/
)
-
[
Qwen2.5-VL-3B-Instruct
](
https://modelscope.cn/models/Qwen/Qwen2.5-VL-3B-Instruct
)
-
[
Qwen2.5-VL-7B-Instruct
](
https://modelscope.cn/models/Qwen/Qwen2.5-VL-7B-Instruct
)
-
[
Qwen2.5-VL-72B-Instruct
](
https://modelscope.cn/models/Qwen/Qwen2.5-VL-72B-Instruct
)
|
**模型名称**
|
**参数量**
|
**推荐 DCU 型号**
|
**最低卡数需求**
|
**下载地址 (Hugging Face)**
|
| --------------------------- | ---------- | ----------------- | ---------------- | ------------------------------------------------------------ |
|
**Qwen2.5-VL-3B-Instruct**
| 3B | K100AI, BW1000 等 | 1 |
[
Hugging Face
](
https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct
)
|
|
**Qwen2.5-VL-7B-Instruct**
| 7B | K100AI, BW1000 等 | 1 |
[
Hugging Face
](
https://huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct
)
|
|
**Qwen2.5-VL-72B-Instruct**
| 72B | K100AI, BW1000 等 | 8 |
[
Hugging Face
](
https://huggingface.co/Qwen/Qwen2.5-VL-72B-Instruct
)
|
## 源码仓库及问题反馈
源码仓库及问题反馈
-
https://developer.sourcefind.cn/codes/modelzoo/Qwen2.5-vl_pytorch
## 参考资料
...
...
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