Skip to main content Crate svod_model Copy item path Source audio Audio preprocessing for ASR: mel features ([mel]) and chunking
([splitter]). blocks Shared ResNet-family building blocks. gigaam GigaAM ASR model. jit resnet ResNet image classification / feature backbone. sentencepiece SentencePiece .model (protobuf) loader. Pieces retain their ▁
(U+2581) prefix on word-initial tokens; consumers replace ▁ with a
space for natural detokenization. silero_vad Silero V5 voice-activity detection. state wespeaker WeSpeaker ResNet34 speaker-embedding model
(pyannote/wespeaker-voxceleb-resnet34-LM). load_state_field Load each named field of $self from the state dict under
<prefix>.<field>. Mirrors state_field! state_field Insert each named field of $self into the state dict under
<prefix>.<field>. Field idents are used verbatim as keys.