{"repo":"NousResearch/hermes-agent","public_id":"110003","display_label":"Sherpa KWS tokenizer type not auto-detected","summary":"A bug where the Sherpa keyword spotting engine hard-codes BPE tokenization causes non-BPE models (e.g. Chinese WenetSpeech) to fail, with accompanying fixes to detect the tokenizer layout from the model directory.","subsystem":"agent","counts":{"members":3,"open_issues":1,"open_issues_distinct":1,"open_prs":2,"closed_issues":0,"merged_prs":0,"attempt_prs":2},"landed_fix":null,"resolution":{"state":"unresolved","by":null},"freshness":{"state":"never_verified","last_verified_at":null,"is_stale":false},"grouping":{"backbone":"STRUCTURAL_SPANNING","link_density":0.6666666666666666},"fix_edges":[],"members":[{"kind":"issue","number":110003,"title":"[Bug]: Wake word: sherpa KWS hard-codes tokens_type=\"bpe\" — any model without bpe.model crashes the engine (Chinese wenetspeech model never starts)","url":null,"state":"open"},{"kind":"pr","number":110087,"title":"fix(tools): support non-BPE Sherpa wake vocabularies","url":null,"state":"open","close_reason":{"refs":[{"kind":"commit","value":"5e9d9ad6b8","number":0},{"kind":"commit","value":"81c627e01f","number":0},{"kind":"pr","value":"#110057","number":110057}],"category":"uncategorized","implication":"needs_llm"},"merged":false},{"kind":"pr","number":124849,"title":"fix(wake-word): pick the sherpa KWS tokenizer layout from the model directory","url":null,"state":"open","close_reason":{"category":"uncategorized","implication":"needs_llm"},"merged":false}],"trust":{"default_tier":"verified","includes_suggested":false}}