-
Notifications
You must be signed in to change notification settings - Fork 121
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
inference model can not see anything useful #10
Comments
the same😅
|
Hi, thanks for trying. Could you try the our inference script instead? There are several mismatches between your implementation and ours. https://github.com/deepseek-ai/DeepSeek-VL2?tab=readme-ov-file#simple-inference-example |
i use "image_placeholder" instead of "image", it works.👍 |
Could you please tell me your output ? |
我部署的是tiny版本,用来测试ocr能力,我试了多张图片,但模型响应的输出一直是下面段话
|
same problem, but pip install transformers==4.38.2 will solve this |
pip install "transformers<4.42" |
I tried everything here, but
|
Seems like downgrading to v4.38.2 works for torch But I can't get it wo work with the latest transformers. I need that support so it can run on Apple MLX. |
So, the model's OCR capability is zero. Can it be understood this way: for multiple different images, it produces fixed and incorrect outputs, is that right? |
No, it does have OCR capabilities. The error that occurred before was due to an incorrect version setting for the transformer, but it has been fixed after modification. |
my env:
cuda 11.8
torch 2.5.1
torchvision 0.19.0
inference code:
output:
The text was updated successfully, but these errors were encountered: