Update README.md

Use unfingerprinted LM prefix cache names

Update LM prefix caches for fingerprint ca92f8f2ce578604

Migrate all system voice caches to version 2 flat snapshot format

Remove nested preset LM prefix cache directory

Flatten preset LM prefix caches

Upload preset voice LM prefix KV caches

Create README.md

Add VAE decode context voice caches

Remove legacy VAE voice caches

Migrate voice caches to feature embeddings

Add included voice caches

Add small Core ML package marker

Add VoxCPM2 config

Upload float16 embed_tokens.npy

Upload projections.mlmodelc

Upload fsq_s4.mlmodelc

Upload audio_vae_decoder_lf4.mlmodelc

Upload feat_encoder.mlmodelc

Upload audio_vae_encoder.mlmodelc

Upload locdit_p4_c4.mlmodelc

Upload residual_lm_fused_multifunction.mlmodelc

Upload base_lm_multifunction.mlmodelc

initial commit