Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

update dependencies and docs to cu118

Browse Source
This commit is contained in:
Lincoln Stein 2023-08-09 13:38:58 -04:00 committed by Kent Keirsey
parent d42b45116f
commit 7bad9bcf53
6 changed files with 41 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -161,7 +161,7 @@ the command `npm install -g yarn` if needed)
_For Windows/Linux with an NVIDIA GPU:_ _For Windows/Linux with an NVIDIA GPU:_
```terminal ```terminal
pip install "InvokeAI[xformers]" --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu117 pip install "InvokeAI[xformers]" --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu118
``` ```
_For Linux with an AMD GPU:_ _For Linux with an AMD GPU:_

2
docs/installation/010_INSTALL_AUTOMATED.md
View File

@ -471,7 +471,7 @@ Then type the following commands:
=== "NVIDIA System" === "NVIDIA System"
```bash ```bash
pip install torch torchvision --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu117 pip install torch torchvision --force-reinstall --extra-index-url https://download.pytorch.org/whl/cu118
pip install xformers pip install xformers
``` ```

6
docs/installation/020_INSTALL_MANUAL.md
View File

@ -148,7 +148,7 @@ manager, please follow these steps:
=== "CUDA (NVidia)" === "CUDA (NVidia)"
```bash ```bash
pip install "InvokeAI[xformers]" --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu117 pip install "InvokeAI[xformers]" --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu118
``` ```
=== "ROCm (AMD)" === "ROCm (AMD)"
@ -312,7 +312,7 @@ installation protocol (important!)
=== "CUDA (NVidia)" === "CUDA (NVidia)"
```bash ```bash
pip install -e .[xformers] --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu117 pip install -e .[xformers] --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu118
``` ```
=== "ROCm (AMD)" === "ROCm (AMD)"
@ -356,7 +356,7 @@ you can do so using this unsupported recipe:
mkdir ~/invokeai mkdir ~/invokeai
conda create -n invokeai python=3.10 conda create -n invokeai python=3.10
conda activate invokeai conda activate invokeai
pip install InvokeAI[xformers] --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu117 pip install InvokeAI[xformers] --use-pep517 --extra-index-url https://download.pytorch.org/whl/cu118
invokeai-configure --root ~/invokeai invokeai-configure --root ~/invokeai
invokeai --root ~/invokeai --web invokeai --root ~/invokeai --web
``` ```

12
docs/installation/030_INSTALL_CUDA_AND_ROCM.md
View File

@ -34,11 +34,11 @@ directly from NVIDIA. **Do not try to install Ubuntu's
nvidia-cuda-toolkit package. It is out of date and will cause nvidia-cuda-toolkit package. It is out of date and will cause
conflicts among the NVIDIA driver and binaries.** conflicts among the NVIDIA driver and binaries.**
Go to [CUDA Toolkit 11.7 Go to [CUDA Toolkit
Downloads](https://developer.nvidia.com/cuda-11-7-0-download-archive), Downloads](https://developer.nvidia.com/cuda-downloads), and use the
and use the target selection wizard to choose your operating system, target selection wizard to choose your operating system, hardware
hardware platform, and preferred installation method (e.g. "local" platform, and preferred installation method (e.g. "local" versus
versus "network"). "network").
This will provide you with a downloadable install file or, depending This will provide you with a downloadable install file or, depending
on your choices, a recipe for downloading and running a install shell on your choices, a recipe for downloading and running a install shell
@ -61,7 +61,7 @@ Runtime Site](https://developer.nvidia.com/nvidia-container-runtime)
When installing torch and torchvision manually with `pip`, remember to provide When installing torch and torchvision manually with `pip`, remember to provide
the argument `--extra-index-url the argument `--extra-index-url
https://download.pytorch.org/whl/cu117` as described in the [Manual https://download.pytorch.org/whl/cu118` as described in the [Manual
Installation Guide](020_INSTALL_MANUAL.md). Installation Guide](020_INSTALL_MANUAL.md).
## :simple-amd: ROCm ## :simple-amd: ROCm

47
docs/installation/070_INSTALL_XFORMERS.md
View File

@ -28,18 +28,21 @@ command line, then just be sure to activate it's virtual environment.
Then run the following three commands: Then run the following three commands:
```sh ```sh
pip install xformers==0.0.16rc425 pip install xformers~=0.0.19
pip install triton pip install triton # WON'T WORK ON WINDOWS
python -m xformers.info output python -m xformers.info output
``` ```
The first command installs `xformers`, the second installs the The first command installs `xformers`, the second installs the
`triton` training accelerator, and the third prints out the `xformers` `triton` training accelerator, and the third prints out the `xformers`
installation status. If all goes well, you'll see a report like the installation status. On Windows, please omit the `triton` package,
which is not available on that platform.
If all goes well, you'll see a report like the
following: following:
```sh ```sh
xFormers 0.0.16rc425 xFormers 0.0.20
memory_efficient_attention.cutlassF: available memory_efficient_attention.cutlassF: available
memory_efficient_attention.cutlassB: available memory_efficient_attention.cutlassB: available
memory_efficient_attention.flshattF: available memory_efficient_attention.flshattF: available
@ -48,22 +51,28 @@ memory_efficient_attention.smallkF: available
memory_efficient_attention.smallkB: available memory_efficient_attention.smallkB: available
memory_efficient_attention.tritonflashattF: available memory_efficient_attention.tritonflashattF: available
memory_efficient_attention.tritonflashattB: available memory_efficient_attention.tritonflashattB: available
indexing.scaled_index_addF: available
indexing.scaled_index_addB: available
indexing.index_select: available
swiglu.dual_gemm_silu: available
swiglu.gemm_fused_operand_sum: available
swiglu.fused.p.cpp: available swiglu.fused.p.cpp: available
is_triton_available: True is_triton_available: True
is_functorch_available: False is_functorch_available: False
pytorch.version: 1.13.1+cu117 pytorch.version: 2.0.1+cu118
pytorch.cuda: available pytorch.cuda: available
gpu.compute_capability: 8.6 gpu.compute_capability: 8.9
gpu.name: NVIDIA RTX A2000 12GB gpu.name: NVIDIA GeForce RTX 4070
build.info: available build.info: available
build.cuda_version: 1107 build.cuda_version: 1108
build.python_version: 3.10.9 build.python_version: 3.10.11
build.torch_version: 1.13.1+cu117 build.torch_version: 2.0.1+cu118
build.env.TORCH_CUDA_ARCH_LIST: 5.0+PTX 6.0 6.1 7.0 7.5 8.0 8.6 build.env.TORCH_CUDA_ARCH_LIST: 5.0+PTX 6.0 6.1 7.0 7.5 8.0 8.6
build.env.XFORMERS_BUILD_TYPE: Release build.env.XFORMERS_BUILD_TYPE: Release
build.env.XFORMERS_ENABLE_DEBUG_ASSERTIONS: None build.env.XFORMERS_ENABLE_DEBUG_ASSERTIONS: None
build.env.NVCC_FLAGS: None build.env.NVCC_FLAGS: None
build.env.XFORMERS_PACKAGE_FROM: wheel-v0.0.16rc425 build.env.XFORMERS_PACKAGE_FROM: wheel-v0.0.20
build.nvcc_version: 11.8.89
source.privacy: open source source.privacy: open source
``` ```
@ -83,14 +92,14 @@ installed from source. These instructions were written for a system
running Ubuntu 22.04, but other Linux distributions should be able to running Ubuntu 22.04, but other Linux distributions should be able to
adapt this recipe. adapt this recipe.
#### 1. Install CUDA Toolkit 11.7 #### 1. Install CUDA Toolkit 11.8
You will need the CUDA developer's toolkit in order to compile and You will need the CUDA developer's toolkit in order to compile and
install xFormers. **Do not try to install Ubuntu's nvidia-cuda-toolkit install xFormers. **Do not try to install Ubuntu's nvidia-cuda-toolkit
package.** It is out of date and will cause conflicts among the NVIDIA package.** It is out of date and will cause conflicts among the NVIDIA
driver and binaries. Instead install the CUDA Toolkit package provided driver and binaries. Instead install the CUDA Toolkit package provided
by NVIDIA itself. Go to [CUDA Toolkit 11.7 by NVIDIA itself. Go to [CUDA Toolkit 11.8
Downloads](https://developer.nvidia.com/cuda-11-7-0-download-archive) Downloads](https://developer.nvidia.com/cuda-11-8-0-download-archive)
and use the target selection wizard to choose your platform and Linux and use the target selection wizard to choose your platform and Linux
distribution. Select an installer type of "runfile (local)" at the distribution. Select an installer type of "runfile (local)" at the
last step. last step.
@ -101,17 +110,17 @@ example, the install script recipe for Ubuntu 22.04 running on a
x86_64 system is: x86_64 system is:
``` ```
wget https://developer.download.nvidia.com/compute/cuda/11.7.0/local_installers/cuda_11.7.0_515.43.04_linux.run wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda_11.8.0_520.61.05_linux.run
sudo sh cuda_11.7.0_515.43.04_linux.run sudo sh cuda_11.8.0_520.61.05_linux.run
``` ```
Rather than cut-and-paste this example, We recommend that you walk Rather than cut-and-paste this example, We recommend that you walk
through the toolkit wizard in order to get the most up to date through the toolkit wizard in order to get the most up to date
installer for your system. installer for your system.
#### 2. Confirm/Install pyTorch 1.13 with CUDA 11.7 support #### 2. Confirm/Install pyTorch 2.01 with CUDA 11.8 support
If you are using InvokeAI 2.3 or higher, these will already be If you are using InvokeAI 3.0.2 or higher, these will already be
installed. If not, you can check whether you have the needed libraries installed. If not, you can check whether you have the needed libraries
using a quick command. Activate the invokeai virtual environment, using a quick command. Activate the invokeai virtual environment,
either by entering the "developer's console", or manually with a either by entering the "developer's console", or manually with a
@ -124,7 +133,7 @@ Then run the command:
python -c 'exec("import torch\nprint(torch.__version__)")' python -c 'exec("import torch\nprint(torch.__version__)")'
``` ```
If it prints __1.13.1+cu117__ you're good. If not, you can install the If it prints __1.13.1+cu118__ you're good. If not, you can install the
most up to date libraries with this command: most up to date libraries with this command:
```sh ```sh

4
installer/lib/installer.py
View File

@ -463,10 +463,10 @@ def get_torch_source() -> (Union[str, None], str):
url = "https://download.pytorch.org/whl/cpu" url = "https://download.pytorch.org/whl/cpu"
if device == "cuda": if device == "cuda":
url = "https://download.pytorch.org/whl/cu117" url = "https://download.pytorch.org/whl/cu118"
optional_modules = "[xformers,onnx-cuda]" optional_modules = "[xformers,onnx-cuda]"
if device == "cuda_and_dml": if device == "cuda_and_dml":
url = "https://download.pytorch.org/whl/cu117" url = "https://download.pytorch.org/whl/cu118"
optional_modules = "[xformers,onnx-directml]" optional_modules = "[xformers,onnx-directml]"
# in all other cases, Torch wheels should be coming from PyPi as of Torch 1.13 # in all other cases, Torch wheels should be coming from PyPi as of Torch 1.13