推理模型时出错

使用的显卡是 intel iris Xe，操作系统是win11
报错时使用的是之前在其他地方训练的模型，但更换至示例模型时也会报错

D:\RVC1006AMD_Intel\RVC1006AMD_Intel1>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897 --dml
2024-11-01 17:45:40 | INFO | configs.config | No supported Nvidia GPU found
2024-11-01 17:45:40 | INFO | configs.config | overwrite v1/32k.json
2024-11-01 17:45:40 | INFO | configs.config | overwrite v1/40k.json
2024-11-01 17:45:40 | INFO | configs.config | overwrite v1/48k.json
2024-11-01 17:45:40 | INFO | configs.config | overwrite v2/48k.json
2024-11-01 17:45:40 | INFO | configs.config | overwrite v2/32k.json
2024-11-01 17:45:40 | INFO | configs.config | overwrite preprocess_per to 3
2024-11-01 17:45:40 | INFO | configs.config | Use DirectML instead
2024-11-01 17:45:40 | INFO | configs.config | Half-precision floating-point: False, device: privateuseone:0
2024-11-01 17:45:42 | INFO | main | Use Language: zh_CN
Running on local URL: http://0.0.0.0:7897
2024-11-01 17:45:51 | INFO | infer.modules.vc.modules | Get sid: 嗲嗲（芸芸）.pth
2024-11-01 17:45:51 | INFO | infer.modules.vc.modules | Loading: assets/weights/嗲嗲（芸芸）.pth
2024-11-01 17:45:52 | INFO | infer.modules.vc.modules | Select index:
2024-11-01 17:48:55 | WARNING | infer.modules.vc.modules | Traceback (most recent call last):
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\modules\vc\modules.py", line 188, in vc_single
audio_opt = self.pipeline.pipeline(
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\modules\vc\pipeline.py", line 375, in pipeline
self.vc(
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\modules\vc\pipeline.py", line 271, in vc
audio1 = (net_g.infer(*arg)[0][0, 0]).data.cpu().float().numpy()
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\models.py", line 791, in infer
m_p, logs_p, x_mask = self.enc_p(phone, pitch, phone_lengths)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\models.py", line 71, in forward
x = self.encoder(x * x_mask, x_mask)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\attentions.py", line 69, in forward
y = attn_layers(x, x, attn_mask)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\attentions.py", line 227, in forward
x, _ = self.attention(q, k, v, mask=attn_mask)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\attentions.py", line 278, in attention
relative_weights = self._absolute_position_to_relative_position(p_attn)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\infer\lib\infer_pack\attentions.py", line 373, in _absolute_position_to_relative_position
x_final = x_flat.view([batch, heads, length, 2 * length])[:, :, :, 1:]
RuntimeError

Traceback (most recent call last):
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "D:\RVC1006AMD_Intel\RVC1006AMD_Intel1\runtime\lib\site-packages\gradio\processing_utils.py", line 219, in convert_to_16_bit_wav
if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype'

Start with infer/modules/vc/pipeline.py at line 271 and infer/lib/infer_pack/attentions.py at line 373, then reproduce the failure using the provided Windows DirectML command and an example model. Trace the RuntimeError before the Gradio NoneType error; done means inference completes without either reported error.