Stop a request in the suffix cache. Args: req_id: Request identifier
(self, req_id: str)
| 108 | spec_logger.debug(f"[SuffixDecoding] Start request {req_id}.") |
| 109 | |
| 110 | def stop_request(self, req_id: str): |
| 111 | """ |
| 112 | Stop a request in the suffix cache. |
| 113 | |
| 114 | Args: |
| 115 | req_id: Request identifier |
| 116 | """ |
| 117 | if req_id in self.suffix_cache.active_requests: |
| 118 | self.suffix_cache.stop_request(req_id) |
| 119 | |
| 120 | # Clean up mappings |
| 121 | if req_id in self.req_id_to_idx: |
| 122 | idx = self.req_id_to_idx[req_id] |
| 123 | |
| 124 | del self.req_id_to_idx[req_id] |
| 125 | if idx in self.idx_to_req_id: |
| 126 | del self.idx_to_req_id[idx] |
| 127 | |
| 128 | spec_logger.debug(f"[SuffixDecoding] Stop request {req_id}.") |
| 129 | |
| 130 | def add_active_response(self, req_id: str, token_ids: list[int]): |
| 131 | """ |