peki12345
peki12345
> I am not sure that it is LORA. Usually LORA is much smaller than 1.32GB Actually, it's a LORA, it fits the definition of LORA, just adjusting the network...
> 数据集不一定要这么多啊，最多就半个小时以内就可以了呀，一般 5-10 分钟就好了。 发件人: peki12345 ***@***.***>日期: 星期三, 2024年2月7日 08:58收件人: RVC-Boss/GPT-SoVITS ***@***.***>抄送: Subscribed ***@***.***>主题: [RVC-Boss/GPT-SoVITS] 感谢大佬开源这个工程，想请教个问题 (Issue #417)我的音源量足够，但是质量可能不太够。这个工程对数据质量要求应该比较高，因为我用3小时普通话纯人声（无手动打标）训出来的模型效果只能算差强人意。在这种情况下，我只能手动一条条打标，筛选出质量高的片段吗，但这个工作量有点大了。想请问各位大佬还有没有什么更省人工的方式。—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving...
> > > Hey @BenjaminIrwin, > > > This is actually quite easily doable. You just need to pass a config parameter that will change the size of your input...
> It's supported, pls see examples/multimodal for more info. Hi, Qwen2-VL can run successfully, but compared to directly import transformers, there is no significant improvement in time consumption and GPU...
> > Nice work! I have a question, why should we concat these two features(LLMs and CLIP) instead of just using LLMs' features, as some other works have done: https://github.com/Kwai-Kolors/Kolors...
请问你是如何用convert_weight_sat2hf脚本转换ckpt的，我2b是可以转的，5b会报错。我似乎该修改某些层的通道配置？但我不知道在哪里修改
> > 请问你是如何用convert_weight_sat2hf脚本转换ckpt的，我2b是可以转的，5b会报错。我似乎该修改某些层的通道配置？但我不知道在哪里修改 > > 这个我们之后更新，转换脚本已经升级了，适配了diffusers 0.30.1 期待更新！
> 我这里稍微改了下 感谢，我用你的修改成功转好了模型，同样遇到了视频变模糊的问题。但我认为这个不是vae导致的，5b模型的vae和2b模型应该没有不同，我更倾向于transformer finetune崩了，但具体原因不清楚，loss也很正常...
> Exiting now because you already have duplicate files, the correct way to download is > > 1. Download latest separately > 2. Open 1 and download that tar.gz separately....
> 超时还是其他原因，我用的wget能正常下载 提示的网络错误，我是直接点的下载按钮，我试试wget