Clip RAM and VRAM to maximum system values

- Prevent a crash in `configure_invokeai` when user has manually set
  RAM or VRAM cache sizes to larger than the size available on system.

- Remove redundant imports from `invokeai.frontend.config.__init__.py`
  which were contributing to circular dependencies.

README.md

@@ -161,7 +161,7 @@ the command `npm install -g yarn` if needed)
     _For Windows/Linux with an NVIDIA GPU:_
 
     ```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:_

docs/installation/010_INSTALL_AUTOMATED.md

@@ -471,7 +471,7 @@ Then type the following commands:
 
 === "NVIDIA System"
     ```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
     ```
 

docs/installation/020_INSTALL_MANUAL.md

@@ -148,7 +148,7 @@ manager, please follow these steps:
     === "CUDA (NVidia)"
 
         ```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)"
@@ -312,7 +312,7 @@ installation protocol (important!)
 
     === "CUDA (NVidia)"
         ```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)"
@@ -356,7 +356,7 @@ you can do so using this unsupported recipe:
 mkdir ~/invokeai
 conda create -n invokeai python=3.10
 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 --root ~/invokeai --web
 ```

docs/installation/030_INSTALL_CUDA_AND_ROCM.md

@@ -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 It is out of date and will cause
 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
 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
 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).
 
 ## :simple-amd: ROCm

docs/installation/070_INSTALL_XFORMERS.md

@@ -28,18 +28,21 @@ command line, then just be sure to activate it's virtual environment.
 Then run the following three commands:
 
 ```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
 ```
 
 The first command installs `xformers`, the second installs the
 `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:
 
 ```sh
-xFormers 0.0.16rc425
+xFormers 0.0.20
 memory_efficient_attention.cutlassF:               available
 memory_efficient_attention.cutlassB:               available
 memory_efficient_attention.flshattF:               available
@@ -48,22 +51,28 @@ memory_efficient_attention.smallkF:                available
 memory_efficient_attention.smallkB:                available
 memory_efficient_attention.tritonflashattF:        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
 is_triton_available:                               True
 is_functorch_available:                            False
-pytorch.version:                                   1.13.1+cu117
+pytorch.version:                                   2.0.1+cu118
 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.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.XFORMERS_BUILD_TYPE:                     Release
 build.env.XFORMERS_ENABLE_DEBUG_ASSERTIONS:        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
 ```
 
@@ -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
 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
 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
 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
 distribution. Select an installer type of "runfile (local)" at the
 last step.
@@ -101,17 +110,17 @@ example, the install script recipe for Ubuntu 22.04 running on a
 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
 through the toolkit wizard in order to get the most up to date
 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
 using a quick command. Activate the invokeai virtual environment,
 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__)")'
 ```
 
-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:
 
 ```sh

installer/lib/installer.py

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

invokeai/backend/install/invokeai_configure.py

@@ -21,7 +21,6 @@ from argparse import Namespace
 from enum import Enum
 from pathlib import Path
 from shutil import get_terminal_size
-from typing import get_type_hints
 from urllib import request
 
 import npyscreen
@@ -399,7 +398,7 @@ Use cursor arrows to make a checkbox selection, and space to toggle.
         self.max_cache_size = self.add_widget_intelligent(
             IntTitleSlider,
             name="RAM cache size (GB). Make this at least large enough to hold a single full model.",
-            value=old_opts.max_cache_size,
+            value=clip(old_opts.max_cache_size, range=(3.0, MAX_RAM)),
             out_of=MAX_RAM,
             lowest=3,
             begin_entry_at=6,
@@ -418,7 +417,7 @@ Use cursor arrows to make a checkbox selection, and space to toggle.
             self.nextrely -= 1
             self.max_vram_cache_size = self.add_widget_intelligent(
                 npyscreen.Slider,
-                value=old_opts.max_vram_cache_size,
+                value=clip(old_opts.max_vram_cache_size, range=(0, MAX_VRAM)),
                 out_of=round(MAX_VRAM * 2) / 2,
                 lowest=0.0,
                 relx=8,
@@ -596,6 +595,16 @@ def default_user_selections(program_opts: Namespace) -> InstallSelections:
     )
 
 
+# -------------------------------------
+def clip(value: float, range: tuple[float, float]) -> float:
+    minimum, maximum = range
+    if value < minimum:
+        value = minimum
+    if value > maximum:
+        value = maximum
+    return value
+
+
 # -------------------------------------
 def initialize_rootdir(root: Path, yes_to_all: bool = False):
     logger.info("Initializing InvokeAI runtime directory")

invokeai/backend/install/migrate_to_3.py

@@ -591,7 +591,6 @@ script, which will perform a full upgrade in place.""",
     # TODO: revisit - don't rely on invokeai.yaml to exist yet!
     dest_is_setup = (dest_root / "models/core").exists() and (dest_root / "databases").exists()
     if not dest_is_setup:
-        import invokeai.frontend.install.invokeai_configure
         from invokeai.backend.install.invokeai_configure import initialize_rootdir
 
         initialize_rootdir(dest_root, True)

invokeai/frontend/install/__init__.py

@@ -1,6 +1,3 @@
 """
 Initialization file for invokeai.frontend.config
 """
-from .invokeai_configure import main as invokeai_configure
-from .invokeai_update import main as invokeai_update
-from .model_install import main as invokeai_model_install

invokeai/frontend/install/invokeai_configure.py

@@ -1,4 +1,4 @@
 """
 Wrapper for invokeai.backend.configure.invokeai_configure
 """
-from ...backend.install.invokeai_configure import main
+from ...backend.install.invokeai_configure import main as invokeai_configure

pyproject.toml

@@ -118,7 +118,7 @@ dependencies = [
 [project.scripts]
 
 # legacy entrypoints; provided for backwards compatibility
-"configure_invokeai.py" = "invokeai.frontend.install:invokeai_configure"
+"configure_invokeai.py" = "invokeai.frontend.install.invokeai_configure:invokeai_configure"
 "textual_inversion.py" = "invokeai.frontend.training:invokeai_textual_inversion"
 
 # shortcut commands to start cli and web
@@ -130,12 +130,12 @@ dependencies = [
 "invokeai-web" = "invokeai.app.api_app:invoke_api"
 
 # full commands
-"invokeai-configure" = "invokeai.frontend.install:invokeai_configure"
+"invokeai-configure" = "invokeai.frontend.install.invokeai_configure:invokeai_configure"
 "invokeai-merge" = "invokeai.frontend.merge:invokeai_merge_diffusers"
 "invokeai-ti" = "invokeai.frontend.training:invokeai_textual_inversion"
-"invokeai-model-install" = "invokeai.frontend.install:invokeai_model_install"
+"invokeai-model-install" = "invokeai.frontend.install.model_install:main"
 "invokeai-migrate3" = "invokeai.backend.install.migrate_to_3:main"
-"invokeai-update" = "invokeai.frontend.install:invokeai_update"
+"invokeai-update" = "invokeai.frontend.install.invokeai_update:main"
 "invokeai-metadata" = "invokeai.frontend.CLI.sd_metadata:print_metadata"
 "invokeai-node-cli" = "invokeai.app.cli_app:invoke_cli"
 "invokeai-node-web" = "invokeai.app.api_app:invoke_api"