This repo hosts the fp16 (half-precision cast, float32 I/O contract; near-lossless, intended for GPU inference - on CPU use the fp32 default) post-training-quantized variant. The default-precision weights auto-download; quantized variants are opt-in: download the .pt and pass its path as the weights argument (the checkpoint's quant manifest rebuilds the quantized structure at load time).

python

from libreyolo import LibreYOLO

m = LibreYOLO("LibreFeyNobgl-matte-fp16.pt")
res = m.predict("product.jpg")
res[0].matte            # (H, W) float alpha in [0, 1]
res[0].save("cut.png")  # transparent-background PNG

Source

Backbone: Swin Transformer v1, Swin-L tier with stage 3 deepened from 18 to 24 blocks (263M parameters). Training data provenance (upstream): not disclosed by Feyn Inc.; this repo redistributes the author's released weights under their Apache-2.0 grant and does not redistribute training data.

Modifications

State-dict metadata-wrap into the LibreYOLO v1.0 checkpoint schema, then post-training quantization with LibreYOLO's quantize API (fp16 (half-precision cast, float32 I/O contract; near-lossless, intended for GPU inference - on CPU use the fp32 default)), stored in the packed finalized format documented in docs/quantization.md and docs/checkpoint_schema.md of the LibreYOLO source repository.

License

Apache License 2.0. See the LICENSE and NOTICE files.

LibreYOLO/LibreFeyNobgl-matte-fp16

LibreFeyNobgl-matte-fp16

Source

Modifications

License

LibreYOLO/
LibreFeyNobgl-matte-fp16