update readme

README.md

@@ -26,9 +26,7 @@ For `SenseVoice`, visit [SenseVoice repo](https://github.com/FunAudioLLM/SenseVo
 
     - [ ] 25hz llama based llm model which supports lora finetune
     - [ ] Support more instruction mode
-    - [ ] Voice conversion
     - [ ] Music generation
-    - [ ] Training script sample based on Mandarin
     - [ ] CosyVoice-500M trained with more multi-lingual data
     - [ ] More...
 
@@ -113,7 +111,7 @@ from cosyvoice.cli.cosyvoice import CosyVoice
 from cosyvoice.utils.file_utils import load_wav
 import torchaudio
 
-cosyvoice = CosyVoice('pretrained_models/CosyVoice-300M-SFT')
+cosyvoice = CosyVoice('pretrained_models/CosyVoice-300M-SFT', load_jit=True, load_onnx=False, fp16=True)
 # sft usage
 print(cosyvoice.list_avaliable_spks())
 # change stream=True for chunk stream inference

cosyvoice/bin/train.py

@@ -74,7 +74,7 @@ def get_args():
                         choices=['model_only', 'model+optimizer'],
                         help='save model/optimizer states')
     parser.add_argument('--timeout',
-                        default=30,
+                        default=60,
                         type=int,
                         help='timeout (in seconds) of cosyvoice_join.')
     parser = deepspeed.add_config_arguments(parser)

examples/libritts/cosyvoice/run.sh

@@ -99,7 +99,6 @@ if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
       --num_workers ${num_workers} \
       --prefetch ${prefetch} \
       --pin_memory \
-      --timeout 300 \
       --deepspeed_config ./conf/ds_stage2.json \
       --deepspeed.save_states model+optimizer
   done