PaddleSpeech
PaddleSpeech copied to clipboard
PaddlePaddle
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation a...
调用 asr = ASRExecutor() current_result = asr(audio_file=audio_path_output, lang="en", model="transformer_librispeech")后gpu内存没有被释放掉，每调用一次，gpu内存都会增加，最后导致gpu崩溃[S2T]XXXX
## General Question 【有手就行】使用你自己的声音做语音合成-这项目有跑起来的吗? https://aistudio.baidu.com/projectdetail/5003396 感觉改了一大堆东西什么用都没有,各种环境报错.无力吐槽
在vits模型文件下的wavenet 文件中的residual_block.py文件中的第127行的下面代码中， x = F.dropout(x, p=self.dropout_rate, training=self.training) 参数training使用的self.training,没有在该py文件的任何地方定义，导致报错。 
我解析出来结果不正确，我用的conformer_online_wenetspeech模型，把原来是16000改成了8000，以下是我的配置文件： `asr_online: model_type: 'conformer_online_wenetspeech' am_model: # the pdmodel file of am static model [optional] am_params: # the pdiparams file of am static model [optional] lang: 'zh' sample_rate: 8000 cfg_path: decode_method:...
成功运行了 [PaddleSpeech-Streaming-ASR-Client](https://github.com/xiaomingnio/PaddleSpeech-Streaming-ASR-Client/tree/db6787e5da75e3948c9120b8af6659a71ec8146f) /websocket_client.py 发现并不支持`pretrained_models.py`中的`conformer_talcs-codeswitch_zh_en-16k`模型， `Can't find "conformer_talcs-codeswitch-zh_en-16k" in resource. Model name must be one of ['conformer_online_wenetspeech-zh-16k', 'conformer_u2pp_online_wenetspeech-zh-16k', 'conformer_online_multicn-zh-16k', 'conformer_online_aishell-zh-16k', 'deepspeech2online_wenetspeech-zh-16k', 'deepspeech2online_aishell-zh-16k']` 找到了`conformer_talcs_application.yaml`中发现`protocol`并不支持`websocket`，`engine_list`也不支持`asr_online`。有大佬知道在哪儿修改吗？
Metadata
Owner
Metadata
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation a...