chlinfeng1997

Why Mini-GTP4 can recognize and output Chinese？

1

I am curious about the reason why this model has the ability to communicate in Chinese.To my knowledge, LLAMA does not support Chinese. Is it because the SharedGPT data used...

Some configuration doubts in train_stage_2.yaml

2

1.Why clip-vit-large-patch14 is selected for clip_model in stage 2 instead of clip-vit-base-patch32 in stage 1？ 2.Why are the configurations of poseguider_checkpoint_path and referencenet_checkpoint_path in stage 2 empty? 3.Why gradient_accumulation_steps is...

After we complete the training of Stage 1 and Stage 2, we try to use demo/animate.py to infer. When loading unet weights, we encounter the problem of missing some weights:...

I noticed that different transformation operations are used for pose and image. For pose: self.cond_transform = transforms.Compose( [ transforms.RandomResizedCrop( self.img_size, scale=self.img_scale, ratio=self.img_ratio, interpolation=transforms.InterpolationMode.BILINEAR, ), transforms.ToTensor(), ] ) For image: self.transform...

A question about Stage 1 training

2

Hello! Thanks for open sourcing your own implementation. During the training phase of Stage 1, is there any precaution to be taken when setting the up_block.3 module in ReferenceNet to...

Why is the temporal layer trained inStage 2 not initialized using the temporal layer weights in AnimateDiff?

5

Thanks for open source！ The original paper mentions using the temporal layer weight of AnimateDiff for initialization, why does it not appear in the code? Looking forward to your reply！

tiktok 下载时出现——响应内容为空，可能是接口失效或者 Cookie 失效，请尝试更新 Cookie

5

感谢作者开源的工具，使用遇到以下问题： 在键入1. 复制粘贴写入 Cookie 后出现 当前 Cookie 已登录 保存配置成功！ 写入 Cookie 成功！ 但是键入4. 终端交互模式出现 缓存数据文件不存在 然后再键入3. 批量下载链接作品(通用) 请输入作品链接: https://www.tiktok.com/@vueltaalmundoenmoto/video/7332886806339357984 出现问题： 响应内容为空，可能是接口失效或者 Cookie 失效，请尝试更新 Cookie 获取作品数据失败 正在尝试第 1 次重试 响应内容为空，可能是接口失效或者 Cookie...

hello，there are 8 gpus on my machine, how to train in parallel?