which pre-trained model is better?

[OutisLi]

there are many models been downloaded , which is better?

789M    ./gsv-v4-pretrained
734M    ./s2Gv3.pth
734M    ./gsv-v4-pretrained/s2Gv4.pth
622M    ./chinese-roberta-wwm-ext-large/pytorch_model.bin
622M    ./chinese-roberta-wwm-ext-large
339M    ./gsv-v2final-pretrained
215M    ./models--nvidia--bigvgan_v2_24khz_100band_256x/bigvgan_generator.pt
215M    ./models--nvidia--bigvgan_v2_24khz_100band_256x
181M    ./chinese-hubert-base/pytorch_model.bin
181M    ./chinese-hubert-base
149M    ./s1v3.ckpt
149M    ./gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
148M    ./s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
126M    ./fast_langdetect/lid.176.bin
126M    ./fast_langdetect
102M    ./s2G488k.pth
102M    ./gsv-v2final-pretrained/s2G2333k.pth
90M     ./s2D488k.pth
90M     ./gsv-v2final-pretrained/s2D2333k.pth
56M     ./gsv-v4-pretrained/vocoder.pth
264K    ./chinese-roberta-wwm-ext-large/tokenizer.json
4.0K    ./models--nvidia--bigvgan_v2_24khz_100band_256x/config.json
4.0K    ./chinese-roberta-wwm-ext-large/config.json
4.0K    ./chinese-hubert-base/preprocessor_config.json
4.0K    ./chinese-hubert-base/config.json
4.0K    ./README.md
4.0K    ./.gitignore

[OutisLi]

the name is so confussing