place input tensors on the same device

#34
by yzmyyff - opened

https://huggingface.co/openbmb/MiniCPM-o-2_6/blob/main/modeling_minicpmo.py#L1345

        text = text.split("<|tts_bos|>")[-1]
        gen_text = text.split("<|tts_eos|>")[0]
        tts_text, tts_token_lens = self.prepare_tts_text(gen_text)
        tts_inputs = self.tts_processor.text_tokenizer.encode(tts_text, add_special_tokens=False)
        tts_input_ids = torch.Tensor(tts_inputs).unsqueeze(0).to("cuda", dtype=torch.long)

tts_input_ids may not be in the same device as the model.

Sign up or log in to comment