software-mansion
/

EfficientNetV2-S-ExecuTorch

Image Classification

Model card Files Files and versions

EfficientNetV2-S-ExecuTorch / qnn

265 MB

Ctrl+K

Ctrl+K

5 contributors

History: 3 commits

msluszniak's picture

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default)

589e267 verified 10 days ago

config.json

6.24 kB
Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago
efficientnet_v2_s_qnn_a16w8_v69.pte

32.8 MB
xet

QNN: re-export a16w8 with MinMaxObserver activations about 1 month ago
efficientnet_v2_s_qnn_a16w8_v73.pte

25.6 MB
xet

QNN: re-export a16w8 with MinMaxObserver activations about 1 month ago
efficientnet_v2_s_qnn_a16w8_v75.pte

25.6 MB
xet

QNN: re-export a16w8 with MinMaxObserver activations about 1 month ago
efficientnet_v2_s_qnn_a16w8_v79.pte

25.6 MB
xet

QNN: re-export a16w8 with MinMaxObserver activations about 1 month ago
efficientnet_v2_s_qnn_a16w8_v81.pte

25.7 MB
xet

QNN: re-export a16w8 with MinMaxObserver activations about 1 month ago
efficientnet_v2_s_qnn_a8w8_v69.pte

32.3 MB
xet

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago
efficientnet_v2_s_qnn_a8w8_v73.pte

24.4 MB
xet

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago
efficientnet_v2_s_qnn_a8w8_v75.pte

24.4 MB
xet

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago
efficientnet_v2_s_qnn_a8w8_v79.pte

24.4 MB
xet

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago
efficientnet_v2_s_qnn_a8w8_v81.pte

24.7 MB
xet

Add the a8w8 QNN variants (1.75x faster for -1.4 top-1; a16w8 stays default) 10 days ago