xldistance

我今天也是这样，发现勾选跳过证书验证就可以用了

Log on to http://localhost:21888 show {"detail":"Not Found"} 

> # This version changes default `-k` of hbbs to `-`, which means setting key on client side is mandatory. If you want to run empty `-k` as before, please...

运行cmake命令时还出现了一个错误， parse op yamls: - D:/workspace/Paddle/paddle/phi/ops/yaml/ops.yaml - D:/workspace/Paddle/paddle/phi/ops/yaml/inconsistent/dygraph_ops.yaml - D:/workspace/Paddle/paddle/phi/ops/yaml/backward.yaml - D:/workspace/Paddle/paddle/phi/ops/yaml/inconsistent/dygraph_backward.yaml - D:/workspace/Paddle/paddle/phi/ops/yaml/fused_ops.yaml - D:/workspace/Paddle/paddle/phi/ops/yaml/legacy/static_ops.yaml validate op yaml: - D:/workspace/Paddle/build/paddle/phi/ops/yaml/ops.parsed.yaml - D:/workspace/Paddle/build/paddle/phi/ops/yaml/backward.parsed.yaml Something Wrong here, this backward op (global_gather_grad)'s...

> 请问问题解决了吗 没有

@lishuai-97 加了＃include编译还是失败 ``` [602/3329] Building CUDA object paddle\phi\CMak...phi_core.dir\kernels\funcs\math_function.cu.ob math_function.cu tmpxft_0000499c_00000000-10_math_function.cudafe1.cpp [603/3329] Building CUDA object paddle\phi\CMak...i_core.dir\kernels\funcs\eigen\broadcast.cu.ob broadcast.cu tmpxft_0000bd04_00000000-10_broadcast.cudafe1.cpp ninja: build stopped: subcommand failed. ```

@lishuai-97 我是windows装的是cuda12.1，然后安装最新的的paddle3.0版本cuda12.3可以正常运行， python -m pip install paddlepaddle-gpu==3.0.0b0 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/，就不折腾自己编译了

Sampling method set to Euler. Schedule type is set to Beta, and the image can be generated normally.

@IsaiChristian You are missing the [ae.safetensors],https://huggingface.co/black-forest-labs/FLUX.1-dev/tree/main

@qw381523977 Recommend you download the bnb-nf4-v2 model so that you don't have to load the vae model additionally