Closed Jia-py closed 6 months ago
All supported models are there: https://huggingface.co/timm
from towhee import pipe, ops, DataCollection
p = (
pipe.input('path')
.map('path', 'img', ops.image_decode.cv2())
.map('img', 'vec', ops.image_embedding.timm(model_name={model_name}))
.output('img', 'vec')
)
Is there an existing issue for this?
What kind of documentation would you like added or changed?
The documentation on the website seems out of date. Some APIs are different now.
I find it challenging to generate embeddings with newly added models (e.g., CVNet).
Why is this needed?
No response
Anything else?
No response