hg38_62celltypes, hg38_33celltypes: .msdref version 3, with a confusion

mm10_41celltypes.msdref: the first mouse deconvolution reference

SHA256SUMS: hg38_wg and both lites rehashed, two withdrawn files dropped

models: new hg38_wg decoder, footprint lites, routing trees withdrawn

README: a pointer to the methscope docs, not a second copy of them

README: bank classifiers, held-out mouse accuracy, classify argument order

SHA256SUMS: add the four bank classifiers

bank classifiers: hg38_celltype_{full,lite}, mm10_brain_{full,lite}

README: hg38_62celltypes.msdref is the primary deconvolution reference

SHA256SUMS: add hg38_62celltypes.msdref

hg38_62celltypes.msdref: Zhou-backbone 62-class deconvolution reference

Remove withdrawn v7 model files

Prepare v7 model release

hg38_celltype: violation framework, 33 classes; .ubjx -> .clfx

SHA256SUMS for the balanced-split model

hg38_wg.updecx: balanced 152/45/10 split, best on the coverage ladder

SHA256SUMS for the curated-split model

hg38_wg.updecx: curated 117/45/45 split (publication reference)

SHA256SUMS: add mm10_wg.updecx

mm10_wg.updecx: coverage-ladder mouse upscale decoder

SHA256SUMS for hg38_wg.updecx v2

hg38_wg.updecx v2: coverage-ladder + scalar-coverage decoder (arm D2A)

Publish SHA256SUMS for digest-anchored fetching (#1)

Upload README.md with huggingface_hub

Upload folder using huggingface_hub

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload hg38_wg.updecx with huggingface_hub

initial commit