facebookresearch / sapiens

High-resolution models for human tasks.
https://about.meta.com/realitylabs/codecavatars/sapiens/
Other
4.53k stars 258 forks source link

Can I use batch input? #142

Closed rabum closed 1 month ago

rabum commented 1 month ago

From what I see the actual inference code is this line: pose_results = inference_topdown(pose_estimator, img, bboxes) The input is the model, one image, and one human detection result, can this have batch input? Like a list of images and bboxes: pose_results = inference_topdown(pose_estimator, img_batches, bboxes_batches)

rawalkhirodkar commented 1 month ago

Yes, you provide batch inputs. However, the code might require minor modifications.

For ready to use batching, please use the lite-inference.