ryugyosoft
/

onw-image

Model card Files Files and versions

onw-image / onw_image

180 kB

Ctrl+K

Ctrl+K

1 contributor

History: 4 commits

ryugyosoft's picture

onw-Image 0.3: Panther Lake ~77 -> ~48 s per 1024 image (fused RMSNorm kernels, FlashAttention-2, async zero-copy blocks; same model files)

6b8ea09 verified 6 days ago

web
onw-Image 0.2: ~47 GB -> ~11 GB install (one weightless graph for all layers, onw NPU kit, xz), tiled VAE, text encoder accuracy fix on the NPU, history scroll fix 7 days ago
__init__.py

133 Bytes
onw-Image 0.3: Panther Lake ~77 -> ~48 s per 1024 image (fused RMSNorm kernels, FlashAttention-2, async zero-copy blocks; same model files) 6 days ago
__main__.py

190 Bytes
onw-Image 0.1a 23 days ago
app.py

16.7 kB
onw-Image 0.1: Windows / Panther Lake (FP8), loop mode, NPU turbo, Lunar Lake residency 22 days ago
backend.py

12.6 kB
onw-Image 0.1a 23 days ago
fetch_model.py

2.52 kB
onw-Image 0.2: ~47 GB -> ~11 GB install (one weightless graph for all layers, onw NPU kit, xz), tiled VAE, text encoder accuracy fix on the NPU, history scroll fix 7 days ago
i18n.json

7.09 kB
onw-Image 0.1: Windows / Panther Lake (FP8), loop mode, NPU turbo, Lunar Lake residency 22 days ago
i18n.py

2.1 kB
onw-Image 0.1a 23 days ago
logo.py

4.32 kB
onw-Image 0.1a 23 days ago
npu_backend.py

15 kB
onw-Image 0.3: Panther Lake ~77 -> ~48 s per 1024 image (fused RMSNorm kernels, FlashAttention-2, async zero-copy blocks; same model files) 6 days ago
server.py

21.8 kB
onw-Image 0.1: Windows / Panther Lake (FP8), loop mode, NPU turbo, Lunar Lake residency 22 days ago