Commit History
-
Add int8 (8-bit k-means-palettized) weight-only tier of both encoders
Verified signatureclap_audio_int8 / clap_text_int8: 8-bit per-tensor k-means palettization of the shipped fp16 weights (activations stay fp16), ~2x smaller. Chosen over linear int8 by measurement (text worst cosine vs fp16: kmeans 0.99873 vs linear 0.99571). int8-vs-fp16 embedding cosine worst: audio 0.99971, text 0.99873. Zero-shot ranking + cross-placement behavior unchanged. coremltools 9.0.02a99c6 -
clapkit T1: LAION CLAP both encoders, fp16 (spectrogram-input audio + RoBERTa text), pre-norm
Verified signature97d631f -
initial commit
Verified signature26bb6b5
