History

Jani Monoses ac51f477eb Add Hiera vision model. (#2382 )		2024-08-01 11:59:22 +02:00
..
README.md	Add Hiera vision model. (#2382 )	2024-08-01 11:59:22 +02:00
main.rs	Add Hiera vision model. (#2382 )	2024-08-01 11:59:22 +02:00

README.md

hiera

Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles This candle implementation uses pre-trained Hiera models from timm for inference. The classification head has been trained on the ImageNet dataset and returns the probabilities for the top-5 classes.

Running an example

$ cargo run --example  hiera --release  -- --image candle-examples/examples/yolo-v8/assets/bike.jpg --which tiny
loaded image Tensor[dims 3, 224, 224; f32]
model built
mountain bike, all-terrain bike, off-roader: 71.15%
unicycle, monocycle     : 7.11%
knee pad                : 4.26%
crash helmet            : 1.48%
moped                   : 1.07%