How can I enable warnings and unavailable? memory_efficient_attention.ckF #1244
Comments
|
I'm not sure I understand the issue. Are you asking about why some of the backends for All the ones starting with The "odd" one is the xformers/xformers/ops/fmha/flash3.py Line 567 in c5c0720 It's quite likely that |
How can I override CUDA_MINIMUM_COMPUTE_CAPABILITY = (9, 0) to CUDA_MINIMUM_COMPUTE_CAPABILITY = (8.9, 0) ? |
|
You would have to modify the source code and reinstall. Also, you would have to set it to |
|
``> You would have to modify the source code and reinstall. Also, you would have to set it to (https://github.com/amDosion/xformers/blob/main/xformers/ops/fmha/flash3.py) I forked the code and modified it to: CUDA_MINIMUM_COMPUTE_CAPABILITY = (8, 9) |
|
BTW, why do you specifically need FlashAttention3? |
It is not particularly necessary, but I just hope that the compilation can be successful, so that if it is needed later, there is no need to recompile. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
❓ Questions and Help
`# ==================================================
🔧 [6.3] Ninja + xformers 编译安装（适配 CUDA 12.8）
==================================================
INSTALL_XFORMERS="${INSTALL_XFORMERS:-true}"
MAIN_REPO_DIR="/app/webui/sd-webui-forge"
XFORMERS_DIR="${MAIN_REPO_DIR}/xformers-src"
TORCH_VER="2.8.0.dev20250326+cu128"
VISION_VER="0.22.0.dev20250326+cu128"
AUDIO_VER="2.6.0.dev20250326+cu128"
TORCH_COMMAND="pip install --pre torch==${TORCH_VER} torchvision==${VISION_VER} torchaudio==${AUDIO_VER} --extra-index-url https://download.pytorch.org/whl/nightly/cu128"
if [[ "$INSTALL_XFORMERS" == "true" ]]; then
echo "⚙️ [6.3] 正在编译并安装 xformers（适配 CUDA 12.8）"
echo "🐍 当前 Python 路径: $(which python)"
echo "🔍 检查 xformers 是否已正确安装..."
if python -c "import xformers" >/dev/null 2>&1 && python -m xformers.info | grep -q "available"; then
echo "✅ 已检测到 xformers 且扩展已启用，版本: $(python -c 'import xformers; print(xformers.version)')"
echo "📦 当前环境无需重新编译 xformers"
fi
echo "📦 检查 PyTorch 是否为指定版本..."
torch_ver=$(pip show torch 2>/dev/null | awk '/^Version:/{print $2}')
vision_ver=$(pip show torchvision 2>/dev/null | awk '/^Version:/{print $2}')
audio_ver=$(pip show torchaudio 2>/dev/null | awk '/^Version:/{print $2}')
if [[ "$torch_ver" != "$TORCH_VER" || "$vision_ver" != "$VISION_VER" || "$audio_ver" != "$AUDIO_VER" ]]; then
echo "🔧 安装指定版本的 PyTorch 组件..."
echo " ➤ $TORCH_COMMAND"
$TORCH_COMMAND && echo " ✅ PyTorch 安装成功" || { echo " ❌ PyTorch 安装失败"; exit 1; }
else
echo " ✅ 已存在所需版本 torch/vision/audio，跳过安装"
fi
echo "📦 检查并安装 pip 构建依赖..."⚠️ 安装 $pkg 失败，请检查网络或 PyPI"
for pkg in pip wheel setuptools cmake ninja; do
ver=$(pip show "$pkg" 2>/dev/null | awk '/^Version:/{print $2}')
if [[ -n "$ver" ]]; then
echo " ✅ $pkg 已安装: v$ver"
else
echo " ❌ $pkg 未安装，尝试安装中..."
pip install "$pkg" --no-cache-dir || echo "
fi
done
✅ 检查系统依赖（仅 root 用户可自动安装）
if [ "$(id -u)" -eq 0 ]; then⚠️ 缺失依赖: ${MISSING[*]}，尝试安装..."⚠️ 当前非 root 用户，跳过系统依赖安装"
echo "🔧 检查系统构建依赖是否已安装..."
MISSING=()
command -v g++ >/dev/null && echo " ✅ g++ 已安装: $(g++ --version | head -n 1)" || MISSING+=("g++")
command -v zip >/dev/null && echo " ✅ zip 已安装: $(zip -v | head -n 1)" || MISSING+=("zip")
command -v unzip >/dev/null && echo " ✅ unzip 已安装: $(unzip -v | head -n 1)" || MISSING+=("unzip")
if [ "${#MISSING[@]}" -eq 0 ]; then
echo "🎉 所有系统构建依赖已满足，无需安装。"
else
echo "
apt-get update && apt-get install -y "${MISSING[@]}"
fi
else
echo "
fi
✅ 准备源码
echo "📁 准备 xformers 源码..."⚠️ 仓库更新失败，跳过"
if [ ! -d "$XFORMERS_DIR/.git" ]; then
echo " - 克隆 xformers 仓库..."
git clone --recursive https://github.com/facebookresearch/xformers.git "$XFORMERS_DIR"
else
echo " - 已存在源码目录，执行 pull..."
cd "$XFORMERS_DIR"
git pull --ff-only || echo "
cd "$MAIN_REPO_DIR"
fi
echo "🔄 初始化子模块（包含 Flash-Attention）..."
cd "$XFORMERS_DIR"
git submodule update --init --recursive || {
echo "❌ 子模块拉取失败，请检查 .gitmodules 或网络连接"
exit 1
}
echo "🔧 开始构建 xformers..."
export XFORMERS_FORCE_CUDA=1
export XFORMERS_BUILD_CPP=1
export TORCH_CUDA_ARCH_LIST="${TORCH_CUDA_ARCH_LIST:-8.9}"
export MAX_JOBS=8
echo " ➤ CUDA 架构: $TORCH_CUDA_ARCH_LIST"
echo " ➤ 并行编译线程数: $MAX_JOBS"
pip install -v -e . --no-build-isolation
build_result=$?
unset XFORMERS_FORCE_CUDA
unset XFORMERS_BUILD_CPP
unset TORCH_CUDA_ARCH_LIST
unset MAX_JOBS
if [ $build_result -ne 0 ]; then
echo "❌ xformers 编译失败，尝试输出诊断信息..."
python -m pip list | grep -E 'torch|xformers|ninja|wheel|cmake|setuptools'
exit 1
fi
echo "✅ xformers 安装完成（已启用 C++ 扩展）"
echo "📋 运行 torch 环境收集..."
python -m torch.utils.collect_env | tee ../torch_env.txt
echo "🧩 运行 xformers.info 检查模块状态..."⚠️ 无法执行 xformers.info"
python -m xformers.info | tee ../xformers_info.txt || echo "
if grep -q "unavailable" ../xformers_info.txt; then⚠️ 以下模块未启用："
echo "
grep "unavailable" ../xformers_info.txt | sed 's/^/ - /'
else
echo "✅ 所有 C++ 扩展模块可用 ✅"
fi
echo "📁 xformers 源码目录: $(realpath "$XFORMERS_DIR")"
echo "🐍 当前 Python: $(which python)"
cd "$MAIN_REPO_DIR"
else
echo "⏭️ [6.3] 跳过 xformers 编译安装（INSTALL_XFORMERS=false）"
fi
log：
[6.3] 正在编译并安装 xformers（适配 CUDA 12.8）
🐍 当前 Python 路径: /app/webui/sd-webui-forge/venv/bin/python
🔍 检查 xformers 是否已正确安装...
✅ 已检测到 xformers 且扩展已启用，版本: 0.0.30+c5c0720c.d20250414
📦 当前环境无需重新编译 xformers
📦 检查 PyTorch 是否为指定版本...
✅ 已存在所需版本 torch/vision/audio，跳过安装
📦 检查并安装 pip 构建依赖...
✅ pip 已安装: v25.0.1
✅ wheel 已安装: v0.45.1
✅ setuptools 已安装: v78.1.0
✅ cmake 已安装: v4.0.0
✅ ninja 已安装: v1.11.1.4
🔧 检查系统构建依赖是否已安装...
✅ g++ 已安装: g++ (GCC) 12.4.0
✅ zip 已安装: Copyright (c) 1990-2008 Info-ZIP - Type 'zip "-L"' for software license.
✅ unzip 已安装: UnZip 6.00 of 20 April 2009, by Debian. Original by Info-ZIP.
🎉 所有系统构建依赖已满足，无需安装。
📁 准备 xformers 源码...
已存在源码目录，执行 pull...
Already up to date.
🔄 初始化子模块（包含 Flash-Attention）...
🔧 开始构建 xformers...
➤ CUDA 架构: 8.9
➤ 并行编译线程数: 8
Using pip 25.0.1 from /app/webui/sd-webui-forge/venv/lib/python3.11/site-packages/pip (python 3.11)
Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/nightly/cu128
Obtaining file:///app/webui/sd-webui-forge/xformers-src
Preparing metadata (setup.py): started
Running command python setup.py egg_info
/app/webui/sd-webui-forge/venv/lib/python3.11/site-packages/setuptools/dist.py:759: SetuptoolsDeprecationWarning: License classifiers are deprecated.
!!
!!
self._finalize_license_expression()
running egg_info
creating /tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info
writing /tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/PKG-INFO
writing dependency_links to /tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/dependency_links.txt
writing top-level names to /tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/top_level.txt
writing manifest file '/tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/SOURCES.txt'
reading manifest file '/tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/SOURCES.txt'
reading manifest template 'MANIFEST.in'
warning: no files found matching 'third_party/flash-attention/version.txt'
adding license file 'LICENSE'
writing manifest file '/tmp/pip-pip-egg-info-9e4nfg6i/xformers.egg-info/SOURCES.txt'
Preparing metadata (setup.py): finished with status 'done'
Installing collected packages: xformers
Attempting uninstall: xformers
Found existing installation: xformers 0.0.30+c5c0720c.d20250414
Uninstalling xformers-0.0.30+c5c0720c.d20250414:
Removing file or directory /app/webui/sd-webui-forge/venv/lib/python3.11/site-packages/xformers.egg-link
Removing pth entries from /app/webui/sd-webui-forge/venv/lib/python3.11/site-packages/easy-install.pth:
Removing entry: /app/webui/sd-webui-forge/xformers-src
Successfully uninstalled xformers-0.0.30+c5c0720c.d20250414
DEPRECATION: Legacy editable install of xformers==0.0.30+c5c0720c.d20250414 from file:///app/webui/sd-webui-forge/xformers-src (setup.py develop) is deprecated. pip 25.1 will enforce this behaviour change. A possible replacement is to add a pyproject.toml or enable --use-pep517, and use setuptools >= 64. If the resulting installation is not behaving as expected, try using --config-settings editable_mode=compat. Please consult the setuptools documentation for more information. Discussion can be found at pypa/pip#11457
Running setup.py develop for xformers
Running command python setup.py develop
/app/webui/sd-webui-forge/venv/lib/python3.11/site-packages/setuptools/dist.py:759: SetuptoolsDeprecationWarning: License classifiers are deprecated.
!!
Successfully installed xformers
✅ xformers 安装完成（已启用 C++ 扩展）
📋 运行 torch 环境收集...
:128: RuntimeWarning: 'torch.utils.collect_env' found in sys.modules after import of package 'torch.utils', but prior to execution of 'torch.utils.collect_env'; this may result in unpredictable behaviour
Collecting environment information...
PyTorch version: 2.8.0.dev20250326+cu128
Is debug build: False
CUDA used to build PyTorch: 12.8
ROCM used to build PyTorch: N/A
OS: Ubuntu 22.04.5 LTS (x86_64)
GCC version: (GCC) 12.4.0
Clang version: 20.1.2 (++20250401083310+e7406753caf3-1
exp120250401203329.93)CMake version: version 4.0.0
Libc version: glibc-2.35
Python version: 3.11.0rc1 (main, Aug 12 2022, 10:02:14) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-6.12.22-Unraid-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: 12.8.93
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: NVIDIA GeForce RTX 4090
Nvidia driver version: 570.133.07
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_engines_precompiled.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_engines_runtime_compiled.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_graph.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_heuristic.so.9.8.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops.so.9.8.0
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 43 bits physical, 48 bits virtual
Byte Order: Little Endian
CPU(s): 32
On-line CPU(s) list: 0-31
Vendor ID: AuthenticAMD
Model name: AMD EPYC 7302 16-Core Processor
CPU family: 23
Model: 49
Thread(s) per core: 2
Core(s) per socket: 16
Socket(s): 1
Stepping: 0
Frequency boost: enabled
CPU max MHz: 3310.5459
CPU min MHz: 1500.0000
BogoMIPS: 5999.79
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es
Virtualization: AMD-V
L1d cache: 512 KiB (16 instances)
L1i cache: 512 KiB (16 instances)
L2 cache: 8 MiB (16 instances)
L3 cache: 128 MiB (8 instances)
NUMA node(s): 1
NUMA node0 CPU(s): 0-31
Vulnerability Gather data sampling: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Mitigation; untrained return thunk; SMT enabled with STIBP protection
Vulnerability Spec rstack overflow: Mitigation; Safe RET
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
Versions of relevant libraries:
⚠️ 以下模块未启用：
[pip3] numpy==2.1.3
[pip3] nvidia-cublas-cu12==12.8.3.14
[pip3] nvidia-cuda-cupti-cu12==12.8.57
[pip3] nvidia-cuda-nvrtc-cu12==12.8.61
[pip3] nvidia-cuda-runtime-cu12==12.8.57
[pip3] nvidia-cudnn-cu12==9.8.0.87
[pip3] nvidia-cufft-cu12==11.3.3.41
[pip3] nvidia-curand-cu12==10.3.9.55
[pip3] nvidia-cusolver-cu12==11.7.2.55
[pip3] nvidia-cusparse-cu12==12.5.7.53
[pip3] nvidia-cusparselt-cu12==0.6.3
[pip3] nvidia-nccl-cu12==2.26.2
[pip3] nvidia-nvjitlink-cu12==12.8.61
[pip3] nvidia-nvtx-cu12==12.8.55
[pip3] open_clip_torch==2.32.0
[pip3] optree==0.15.0
[pip3] pytorch-lightning==2.5.1
[pip3] pytorch-triton==3.3.0+git96316ce5
[pip3] torch==2.8.0.dev20250326+cu128
[pip3] torchaudio==2.6.0.dev20250326+cu128
[pip3] torchdiffeq==0.2.5
[pip3] torchmetrics==1.7.1
[pip3] torchsde==0.2.6
[pip3] torchvision==0.22.0.dev20250326+cu128
[conda] Could not collect
🧩 运行 xformers.info 检查模块状态...
xFormers 0.0.30+c5c0720c.d20250414
memory_efficient_attention.ckF: unavailable
memory_efficient_attention.ckB: unavailable
memory_efficient_attention.ck_decoderF: unavailable
memory_efficient_attention.ck_splitKF: unavailable
memory_efficient_attention.cutlassF-pt: available
memory_efficient_attention.cutlassB-pt: available
[email protected]: available
[email protected]: available
[email protected]: unavailable
[email protected]: unavailable
memory_efficient_attention.triton_splitKF: available
indexing.scaled_index_addF: available
indexing.scaled_index_addB: available
indexing.index_select: available
sp24.sparse24_sparsify_both_ways: available
sp24.sparse24_apply: available
sp24.sparse24_apply_dense_output: available
sp24._sparse24_gemm: available
[email protected]: available
[email protected]: available
swiglu.dual_gemm_silu: available
swiglu.gemm_fused_operand_sum: available
swiglu.fused.p.cpp: available
is_triton_available: True
pytorch.version: 2.8.0.dev20250326+cu128
pytorch.cuda: available
gpu.compute_capability: 8.9
gpu.name: NVIDIA GeForce RTX 4090
dcgm_profiler: unavailable
build.info: available
build.cuda_version: 1208
build.hip_version: None
build.python_version: 3.11.0rc1
build.torch_version: 2.8.0.dev20250326+cu128
build.env.TORCH_CUDA_ARCH_LIST: 8.9
build.env.PYTORCH_ROCM_ARCH: None
build.env.XFORMERS_BUILD_TYPE: None
build.env.XFORMERS_ENABLE_DEBUG_ASSERTIONS: None
build.env.NVCC_FLAGS: None
build.env.XFORMERS_PACKAGE_FROM: None
build.nvcc_version: 12.8.93
source.privacy: open source
- memory_efficient_attention.ckF: unavailable
- memory_efficient_attention.ckB: unavailable
- memory_efficient_attention.ck_decoderF: unavailable
- memory_efficient_attention.ck_splitKF: unavailable
- [email protected]: unavailable
- [email protected]: unavailable
- dcgm_profiler: unavailable
📁 xformers 源码目录: /app/webui/sd-webui-forge/xformers-src`
How can I enable warnings and unavailable?
The text was updated successfully, but these errors were encountered: