abhishekchohan/flux2-klein-4b-fp8

Delete text_encoder/model.safetensors

Add side-by-side BF16-vs-FP8 samples (LFS) and embed in model card

test updateFile upsert

test addFile

Add side-by-side BF16-vs-FP8 samples and embed them in the model card

Model card: richer 16-pair metrics, PickScore/CLIP parity, efficiency table, eval protocol; drop private-repo provenance

Update README.md

Update model card with measured quality metrics

FLUX.2 Klein FP8-quantized pipeline (torchao float8wo, transformer + text encoder)

FLUX.2 Klein FP8-quantized pipeline (torchao float8wo)

initial commit