Using RX6650XT, rocm5.4.2 reports an error when running stable diffusion

[14790897]

python launch.py --precision full --no-half
Python 3.9.17 (main, Jul 5 2023, 20:41:20)
[GCC 11.2.0]
Version: v1.4.1
Commit hash: f865d3e11647dfd6c7b2cdf90dde24680e58acd8
Installing requirements
Launching Web UI with arguments: --precision full --no-half
No module 'xformers'. Proceeding without it.
Loading weights [18ed2b6c48] from /home/ubuntu/sd/stable-diffusion-webui/models/Stable-diffusion/xxmix9realistic_v40.safetensors
preload_extensions_git_metadata for 7 extensions took 0.00s
Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch().
Startup time: 2.9s (import torch: 0.9s, import gradio: 0.5s, import ldm: 0.2s, other imports: 0.5s, load scripts: 0.3s, create ui: 0.4s, gradio launch: 0.1s).
MIOpen(HIP): Error [Compile] 'hiprtcCompileProgram(prog.get(), c_options.size(), c_options.data())' naive_conv.cpp: HIPRTC_ERROR_COMPILATION (6)
MIOpen(HIP): Error [BuildHip] HIPRTC status = HIPRTC_ERROR_COMPILATION (6), source file: naive_conv.cpp
MIOpen(HIP): Warning [BuildHip] /tmp/comgr-8f15a2/input/naive_conv.cpp:39:10: fatal error: 'limits' file not found
#include // std::numeric_limits
^~~~~~~~
1 error generated when compiling for gfx1030.
terminate called after throwing an instance of 'miopen::Exception'
what(): /long_pathname_so_that_rpms_can_package_the_debug_info/data/driver/MLOpen/src/hipoc/hipoc_program.cpp:304: Code object build failed. Source: naive_conv.cpp
Aborted (core dumped)

[14790897]

pytorch can run a bit：

import torch

print(torch.version)
2.0.1+rocm5.4.2
if torch.cuda.is_available():
... x = torch.rand(5, 3).cuda()
... print(x)
... else:
... print("No AMD GPU available")
...
tensor([[0.8301, 0.5985, 0.4911],
[0.8105, 0.9210, 0.2274],
[0.7984, 0.7864, 0.3693],
[0.8469, 0.1225, 0.1067],
[0.9466, 0.5537, 0.4603]], device='cuda:0')

then, when I train the neural network
(sd) ubuntu@ubuntu-System-Product-Name:~/Documents/git-program/neural-net$ python mni.py
MIOpen(HIP): Error [Compile] 'hiprtcCompileProgram(prog.get(), c_options.size(), c_options.data())' naive_conv.cpp: HIPRTC_ERROR_COMPILATION (6)
MIOpen(HIP): Error [BuildHip] HIPRTC status = HIPRTC_ERROR_COMPILATION (6), source file: naive_conv.cpp
MIOpen(HIP): Warning [BuildHip] /tmp/comgr-d09741/input/naive_conv.cpp:39:10: fatal error: 'limits' file not found
#include // std::numeric_limits
^~~~~~~~
1 error generated when compiling for gfx1032.
terminate called after throwing an instance of 'miopen::Exception'
what(): /long_pathname_so_that_rpms_can_package_the_debug_info/data/driver/MLOpen/src/hipoc/hipoc_program.cpp:304: Code object build failed. Source: naive_conv.cpp
Aborted (core dumped)

[14790897]

When I use rocm5.6.0, I enter the webpage of sd webui, but there is a similar error as before when generating the picture, which is 'limits' file not found'
error information:
DiffusionWrapper has 859.52 M params.
Downloading (…)olve/main/vocab.json: 100%|████| 961k/961k [00:01<00:00, 830kB/s]
Downloading (…)olve/main/merges.txt: 100%|███| 525k/525k [00:00<00:00, 2.86MB/s]
Downloading (…)cial_tokens_map.json: 100%|█████| 389/389 [00:00<00:00, 2.54MB/s]
Downloading (…)okenizer_config.json: 100%|█████| 905/905 [00:00<00:00, 5.86MB/s]
Downloading (…)lve/main/config.json: 100%|█| 4.52k/4.52k [00:00<00:00, 23.5MB/s]
Applying attention optimization: Doggettx... done.
Textual inversion embeddings loaded(0):
Model loaded in 12.7s (calculate hash: 4.6s, load weights from disk: 0.1s, create model: 6.0s, apply weights to model: 0.9s, apply half(): 0.4s, move model to device: 0.4s, load textual inversion embeddings: 0.1s, calculate empty prompt: 0.2s).
Calculating sha256 for /home/ubuntu/sd/stable-diffusion-webui/models/Stable-diffusion/xxmix9realistic_v40.safetensors: 18ed2b6c48fda400330e5dec9e6f4d714ef664869ea8d4021f12ba699b31da06
Loading weights [18ed2b6c48] from /home/ubuntu/sd/stable-diffusion-webui/models/Stable-diffusion/xxmix9realistic_v40.safetensors
Applying attention optimization: Doggettx... done.
Weights loaded in 4.2s (calculate hash: 2.3s, load weights from disk: 0.4s, apply weights to model: 1.1s, move model to device: 0.4s).
0%| | 0/20 [00:00<?, ?it/s]MIOpen(HIP): Error [Compile] 'hiprtcCompileProgram(prog.get(), c_options.size(), c_options.data())' convolution_forward_implicit_gemm_v6r1_dlops_nchw_kcyx_nkhw.cpp: HIPRTC_ERROR_COMPILATION (6)
MIOpen(HIP): Error [BuildHip] HIPRTC status = HIPRTC_ERROR_COMPILATION (6), source file: convolution_forward_implicit_gemm_v6r1_dlops_nchw_kcyx_nkhw.cpp
MIOpen(HIP): Warning [BuildHip] In file included from /tmp/comgr-58dcf8/input/convolution_forward_implicit_gemm_v6r1_dlops_nchw_kcyx_nkhw.cpp:1:
In file included from /tmp/comgr-58dcf8/include/common_header.hpp:10:
/tmp/comgr-58dcf8/include/data_type.hpp:14:10: fatal error: 'limits' file not found
#include // std::numeric_limits
^~~~~~~~
1 error generated when compiling for gfx1030.
terminate called after throwing an instance of 'miopen::Exception'
what(): /MIOpen/src/hipoc/hipoc_program.cpp:304: Code object build failed. Source: convolution_forward_implicit_gemm_v6r1_dlops_nchw_kcyx_nkhw.cpp
./webui.sh: line 241: 16653 Aborted (core dumped) "${python_cmd}" "${LAUNCH_SCRIPT}" "$@"

[14790897]

'limits' file not found' is the problem we need to solve

[14790897]

I have spent four days on installing rocm and pytorch, can you guys help me?

[xuhuisheng]

pytorch 1.13 should fine.
there must be bug on pytorch 2.x with rocm.

[14790897]

just run 'sudo apt install libstdc++-12-dev'

[14790897]

https://are-we-gfx1100-yet.github.io/post/a1111-webui/#fatal-error-limits-file-not-found