Method _gen

lmdeploy/pipeline.py:219–231

()

                                      **kwargs)

        def _gen():
            resp = None
            try:
                for out in generator:
                    resp = resp.extend(out) if resp else out
                    yield out
            except:  # noqa
                self._run(coro=session.async_abort())
                raise
            else:
                session.response = resp
                session.step += resp.generate_token_len + resp.input_token_len
                session.history.append((session.prompt, resp.text))

        if stream_response:
            return _gen()