5
                            我在推理时使用了stream_return=True,但是报错:The following model_kwargs are not used by the model: ['stream_return', 'more_segment_before'] (note: typos in the generate arguments will also show up in this list)
我在推理时使用了stream_return=True,但是报错:The following model_kwargs are not used by the model: ['stream_return', 'more_segment_before'] (note: typos in the generate arguments will also show up in this list)
我注释掉了indextts/gpt/transformers_generation_utils.py中1973行的self._validate_model_kwargs(model_kwargs.copy()) 可以运行了但是返回结果有问题: wavs = tts.infer( stream_return=True, # yield each item in wavs spk_audio_prompt='29693316630-1-192_00_00_10-00_00_26.mp3', output_path="gen.wav", text=text,
emo_audio_prompt="examples/voice_05.wav",    # emo_alpha=0.9,
    verbose=True,
    max_text_tokens_per_segment=120,
    more_segment_before=80    ) for wav in wavs: print('------', wav) 这段代码的输出结果是: ------ g ------ e ------ n ------ . ------ w ------ a ------ v
我也遇到了这个问题
拉了最新代码,好了