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

Merge branch 'development' into development

Browse Source
This commit is contained in:
Peter Baylies 2022-09-26 09:21:15 -04:00 committed by GitHub
commit be99d5a4bd
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
25 changed files with 1208 additions and 325 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
README.md
View File

@ -2,6 +2,15 @@
# InvokeAI: A Stable Diffusion Toolkit # InvokeAI: A Stable Diffusion Toolkit
_Note: This fork is rapidly evolving. Please use the
[Issues](https://github.com/invoke-ai/InvokeAI/issues) tab to
report bugs and make feature requests. Be sure to use the provided
templates. They will help aid diagnose issues faster._
_This repository was formally known as lstein/stable-diffusion_
# **Table of Contents**
![project logo](docs/assets/logo.png) ![project logo](docs/assets/logo.png)
[![discord badge]][discord link] [![discord badge]][discord link]

73
backend/server.py
View File

@ -103,6 +103,8 @@ socketio = SocketIO(
engineio_logger=engineio_logger, engineio_logger=engineio_logger,
max_http_buffer_size=max_http_buffer_size, max_http_buffer_size=max_http_buffer_size,
cors_allowed_origins=cors_allowed_origins, cors_allowed_origins=cors_allowed_origins,
ping_interval=(50, 50),
ping_timeout=60,
) )
@ -186,17 +188,50 @@ def handle_request_capabilities():
socketio.emit("systemConfig", config) socketio.emit("systemConfig", config)
@socketio.on("requestAllImages") @socketio.on("requestImages")
def handle_request_all_images(): def handle_request_images(page=1, offset=0, last_mtime=None):
print(f">> All images requested") chunk_size = 50
paths = list(filter(os.path.isfile, glob.glob(result_path + "*.png")))
paths.sort(key=lambda x: os.path.getmtime(x)) if last_mtime:
print(f">> Latest images requested")
else:
print(
f">> Page {page} of images requested (page size {chunk_size} offset {offset})"
)
paths = glob.glob(os.path.join(result_path, "*.png"))
sorted_paths = sorted(paths, key=lambda x: os.path.getmtime(x), reverse=True)
if last_mtime:
image_paths = filter(lambda x: os.path.getmtime(x) > last_mtime, sorted_paths)
else:
image_paths = sorted_paths[
slice(chunk_size * (page - 1) + offset, chunk_size * page + offset)
]
page = page + 1
image_array = [] image_array = []
for path in paths:
for path in image_paths:
metadata = retrieve_metadata(path) metadata = retrieve_metadata(path)
image_array.append({"url": path, "metadata": metadata["sd-metadata"]}) image_array.append(
socketio.emit("galleryImages", {"images": image_array}) {
eventlet.sleep(0) "url": path,
"mtime": os.path.getmtime(path),
"metadata": metadata["sd-metadata"],
}
)
socketio.emit(
"galleryImages",
{
"images": image_array,
"nextPage": page,
"offset": offset,
"onlyNewImages": True if last_mtime else False,
},
)
@socketio.on("generateImage") @socketio.on("generateImage")
@ -275,6 +310,7 @@ def handle_run_esrgan_event(original_image, esrgan_parameters):
"esrganResult", "esrganResult",
{ {
"url": os.path.relpath(path), "url": os.path.relpath(path),
"mtime": os.path.getmtime(path),
"metadata": metadata, "metadata": metadata,
}, },
) )
@ -343,6 +379,7 @@ def handle_run_gfpgan_event(original_image, gfpgan_parameters):
"gfpganResult", "gfpganResult",
{ {
"url": os.path.relpath(path), "url": os.path.relpath(path),
"mtime": os.path.mtime(path),
"metadata": metadata, "metadata": metadata,
}, },
) )
@ -642,7 +679,11 @@ def generate_images(generation_parameters, esrgan_parameters, gfpgan_parameters)
step_index += 1 step_index += 1
socketio.emit( socketio.emit(
"intermediateResult", "intermediateResult",
{"url": os.path.relpath(path), "metadata": generation_parameters}, {
"url": os.path.relpath(path),
"mtime": os.path.getmtime(path),
"metadata": generation_parameters,
},
) )
socketio.emit("progressUpdate", progress) socketio.emit("progressUpdate", progress)
eventlet.sleep(0) eventlet.sleep(0)
@ -670,6 +711,11 @@ def generate_images(generation_parameters, esrgan_parameters, gfpgan_parameters)
first_seed = first_seed or seed first_seed = first_seed or seed
this_variation = [[seed, all_parameters["variation_amount"]]] this_variation = [[seed, all_parameters["variation_amount"]]]
all_parameters["with_variations"] = prior_variations + this_variation all_parameters["with_variations"] = prior_variations + this_variation
all_parameters["seed"] = first_seed
elif ("with_variations" in all_parameters):
all_parameters["seed"] = first_seed
else:
all_parameters["seed"] = seed
if esrgan_parameters: if esrgan_parameters:
progress["currentStatus"] = "Upscaling" progress["currentStatus"] = "Upscaling"
@ -702,7 +748,6 @@ def generate_images(generation_parameters, esrgan_parameters, gfpgan_parameters)
postprocessing = True postprocessing = True
all_parameters["gfpgan_strength"] = gfpgan_parameters["strength"] all_parameters["gfpgan_strength"] = gfpgan_parameters["strength"]
all_parameters["seed"] = first_seed
progress["currentStatus"] = "Saving image" progress["currentStatus"] = "Saving image"
socketio.emit("progressUpdate", progress) socketio.emit("progressUpdate", progress)
eventlet.sleep(0) eventlet.sleep(0)
@ -735,7 +780,11 @@ def generate_images(generation_parameters, esrgan_parameters, gfpgan_parameters)
socketio.emit( socketio.emit(
"generationResult", "generationResult",
{"url": os.path.relpath(path), "metadata": metadata}, {
"url": os.path.relpath(path),
"mtime": os.path.getmtime(path),
"metadata": metadata,
},
) )
eventlet.sleep(0) eventlet.sleep(0)

224
docs/features/CLI.md
View File

@ -103,157 +103,141 @@ overridden on a per-prompt basis (see [List of prompt arguments]
These arguments are deprecated but still work: These arguments are deprecated but still work:
<figure markdown>
| Argument | Shortcut | Default | Description | | Argument | Shortcut | Default | Description |
| ------------------ | -------- | ------- | --------------------------------------------------------------- | |--------------------|------------|---------------------|--------------|
| `--weights <path>` | | `None` | Pth to weights file; use `--model stable-diffusion-1.4` instead | | --weights <path> | | None | Pth to weights file; use `--model stable-diffusion-1.4` instead |
| `--laion400m` | `-l` | `False` | Use older LAION400m weights; use `--model=laion400m` instead | | --laion400m | -l | False | Use older LAION400m weights; use `--model=laion400m` instead |
</figure> **A note on path names:** On Windows systems, you may run into
problems when passing the dream script standard backslashed path
names because the Python interpreter treats "\" as an escape.
You can either double your slashes (ick): C:\\\\path\\\\to\\\\my\\\\file, or
use Linux/Mac style forward slashes (better): C:/path/to/my/file.
!!! note ## List of prompt arguments
On Windows systems, you may run into problems when passing the dream script standard backslashed After the dream.py script initializes, it will present you with a
path names because the Python interpreter treats `\` as an escape. You can either double your **dream>** prompt. Here you can enter information to generate images
slashes (ick): `C:\\path\\to\\my\\file`, or use Linux/Mac style forward slashes (better): from text (txt2img), to embellish an existing image or sketch
`C:/path/to/my/file`. (img2img), or to selectively alter chosen regions of the image
(inpainting).
### List of prompt arguments ### This is an example of txt2img:
After the `dream.py` script initializes, it will present you with a **`dream>`** ~~~~
prompt. Here you can enter information to generate images from text (txt2img), dream> waterfall and rainbow -W640 -H480
to embellish an existing image or sketch (img2img), or to selectively alter ~~~~
chosen regions of the image (inpainting).
#### txt2img This will create the requested image with the dimensions 640 (width)
and 480 (height).
!!! example Here are the dream> command that apply to txt2img:
```bash | Argument | Shortcut | Default | Description |
dream> "waterfall and rainbow" -W640 -H480 |--------------------|------------|---------------------|--------------|
``` | "my prompt" | | | Text prompt to use. The quotation marks are optional. |
| --width <int> | -W<int> | 512 | Width of generated image |
| --height <int> | -H<int> | 512 | Height of generated image |
| --iterations <int> | -n<int> | 1 | How many images to generate from this prompt |
| --steps <int> | -s<int> | 50 | How many steps of refinement to apply |
| --cfg_scale <float>| -C<float> | 7.5 | How hard to try to match the prompt to the generated image; any number greater than 1.0 works, but the useful range is roughly 5.0 to 20.0 |
| --seed <int> | -S<int> | None | Set the random seed for the next series of images. This can be used to recreate an image generated previously.|
| --sampler <sampler>| -A<sampler>| k_lms | Sampler to use. Use -h to get list of available samplers. |
| --grid | -g | False | Turn on grid mode to return a single image combining all the images generated by this prompt |
| --individual | -i | True | Turn off grid mode (deprecated; leave off --grid instead) |
| --outdir <path> | -o<path> | outputs/img_samples | Temporarily change the location of these images |
| --seamless | | False | Activate seamless tiling for interesting effects |
| --log_tokenization | -t | False | Display a color-coded list of the parsed tokens derived from the prompt |
| --skip_normalization| -x | False | Weighted subprompts will not be normalized. See [Weighted Prompts](./OTHER.md#weighted-prompts) |
| --upscale <int> <float> | -U <int> <float> | -U 1 0.75| Upscale image by magnification factor (2, 4), and set strength of upscaling (0.0-1.0). If strength not set, will default to 0.75. |
| --gfpgan_strength <float> | -G <float> | -G0 | Fix faces using the GFPGAN algorithm; argument indicates how hard the algorithm should try (0.0-1.0) |
| --save_original | -save_orig| False | When upscaling or fixing faces, this will cause the original image to be saved rather than replaced. |
| --variation <float> |-v<float>| 0.0 | Add a bit of noise (0.0=none, 1.0=high) to the image in order to generate a series of variations. Usually used in combination with -S<seed> and -n<int> to generate a series a riffs on a starting image. See [Variations](./VARIATIONS.md). |
| --with_variations <pattern> | -V<pattern>| None | Combine two or more variations. See [Variations](./VARIATIONS.md) for now to use this. |
This will create the requested image with the dimensions 640 (width) and 480 (height). Note that the width and height of the image must be multiples of
64. You can provide different values, but they will be rounded down to
the nearest multiple of 64.
Those are the `dream` commands that apply to txt2img:
| Argument <img width="680" align="right"/> | Shortcut <img width="420" align="right"/> | Default <img width="480" align="right"/> | Description | ### This is an example of img2img:
| ----------------------------------------- | ----------------------------------------- | ---------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `"my prompt"` | | | Text prompt to use. The quotation marks are optional. |
| `--width <int>` | `-W<int>` | `512` | Width of generated image |
| `--height <int>` | `-H<int>` | `512` | Height of generated image |
| `--iterations <int>` | `-n<int>` | `1` | How many images to generate from this prompt |
| `--steps <int>` | `-s<int>` | `50` | How many steps of refinement to apply |
| `--cfg_scale <float>` | `-C<float>` | `7.5` | How hard to try to match the prompt to the generated image; any number greater than 0.0 works, but the useful range is roughly 5.0 to 20.0 |
| `--seed <int>` | `-S<int>` | `None` | Set the random seed for the next series of images. This can be used to recreate an image generated previously. |
| `--sampler <sampler>` | `-A<sampler>` | `k_lms` | Sampler to use. Use `-h` to get list of available samplers. |
| `--grid` | `-g` | `False` | Turn on grid mode to return a single image combining all the images generated by this prompt |
| `--individual` | `-i` | `True` | Turn off grid mode (deprecated; leave off `--grid` instead) |
| `--outdir <path>` | `-o<path>` | `outputs/img_samples` | Temporarily change the location of these images |
| `--seamless` | | `False` | Activate seamless tiling for interesting effects |
| `--log_tokenization` | `-t` | `False` | Display a color-coded list of the parsed tokens derived from the prompt |
| `--skip_normalization` | `-x` | `False` | Weighted subprompts will not be normalized. See [Weighted Prompts](./OTHER.md#weighted-prompts) |
| `--upscale <int> <float>` | `-U <int> <float>` | `-U 1 0.75` | Upscale image by magnification factor (2, 4), and set strength of upscaling (0.0-1.0). If strength not set, will default to 0.75. |
| `--gfpgan_strength <float>` | `-G <float>` | `-G0` | Fix faces using the GFPGAN algorithm; argument indicates how hard the algorithm should try (0.0-1.0) |
| `--save_original` | `-save_orig` | `False` | When upscaling or fixing faces, this will cause the original image to be saved rather than replaced. |
| `--variation <float>` | `-v<float>` | `0.0` | Add a bit of noise (0.0=none, 1.0=high) to the image in order to generate a series of variations. Usually used in combination with `-S<seed>` and `-n<int>` to generate a series a riffs on a starting image. See [Variations](./VARIATIONS.md). |
| `--with_variations <pattern>` | `-V<pattern>` | `None` | Combine two or more variations. See [Variations](./VARIATIONS.md) for now to use this. |
!!! note ~~~~
dream> waterfall and rainbow -I./vacation-photo.png -W640 -H480 --fit
~~~~
The width and height of the image must be multiples of 64. You can provide different This will modify the indicated vacation photograph by making it more
values, but they will be rounded down to the nearest multiple of 64. like the prompt. Results will vary greatly depending on what is in the
image. We also ask to --fit the image into a box no bigger than
640x480. Otherwise the image size will be identical to the provided
photo and you may run out of memory if it is large.
#### img2img In addition to the command-line options recognized by txt2img, img2img
accepts additional options:
!!! example | Argument | Shortcut | Default | Description |
|--------------------|------------|---------------------|--------------|
| --init_img <path> | -I<path> | None | Path to the initialization image |
| --fit | -F | False | Scale the image to fit into the specified -H and -W dimensions |
| --strength <float> | -s<float> | 0.75 | How hard to try to match the prompt to the initial image. Ranges from 0.0-0.99, with higher values replacing the initial image completely.|
```bash ### This is an example of inpainting:
dream> "waterfall and rainbow" -I./vacation-photo.png -W640 -H480 --fit
```
This will modify the indicated vacation photograph by making it more like the prompt. Results will ~~~~
vary greatly depending on what is in the image. We also ask to --fit the image into a box no bigger dream> waterfall and rainbow -I./vacation-photo.png -M./vacation-mask.png -W640 -H480 --fit
than 640x480. Otherwise the image size will be identical to the provided photo and you may run out ~~~~
of memory if it is large.
Repeated chaining of img2img on an image can result in significant color shifts This will do the same thing as img2img, but image alterations will
in the output, especially if run with lower strength. Color correction can be only occur within transparent areas defined by the mask file specified
run against a reference image to fix this issue. Use the original input image to by -M. You may also supply just a single initial image with the areas
the chain as the the reference image for each step in the chain. to overpaint made transparent, but you must be careful not to destroy
the pixels underneath when you create the transparent areas. See
[Inpainting](./INPAINTING.md) for details.
In addition to the command-line options recognized by txt2img, img2img accepts inpainting accepts all the arguments used for txt2img and img2img, as
additional options: well as the --mask (-M) argument:
| Argument <img width="160" align="right"/> | Shortcut | Default | Description | | Argument | Shortcut | Default | Description |
| ----------------------------------------- | ----------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------ | |--------------------|------------|---------------------|--------------|
| `--init_img <path>` | `-I<path>` | `None` | Path to the initialization image | | --init_mask <path> | -M<path> | None |Path to an image the same size as the initial_image, with areas for inpainting made transparent.|
| `--init_color <path>` | | `None` | Path to reference image for color correction |
| `--fit` | `-F` | `False` | Scale the image to fit into the specified -H and -W dimensions |
| `--strength <float>` | `-f<float>` | `0.75` | How hard to try to match the prompt to the initial image. Ranges from 0.0-0.99, with higher values replacing the initial image completely. |
#### Inpainting
!!! example # Command-line editing and completion
```bash If you are on a Macintosh or Linux machine, the command-line offers
dream> "waterfall and rainbow" -I./vacation-photo.png -M./vacation-mask.png -W640 -H480 --fit convenient history tracking, editing, and command completion.
```
This will do the same thing as img2img, but image alterations will only occur within transparent - To scroll through previous commands and potentially edit/reuse them, use the up and down cursor keys.
areas defined by the mask file specified by `-M`. You may also supply just a single initial image with - To edit the current command, use the left and right cursor keys to position the cursor, and then backspace, delete or insert characters.
the areas to overpaint made transparent, but you must be careful not to destroy the pixels - To move to the very beginning of the command, type CTRL-A (or command-A on the Mac)
underneath when you create the transparent areas. See [Inpainting](./INPAINTING.md) for details. - To move to the end of the command, type CTRL-E.
- To cut a section of the command, position the cursor where you want to start cutting and type CTRL-K.
- To paste a cut section back in, position the cursor where you want to paste, and type CTRL-Y
Inpainting accepts all the arguments used for txt2img and img2img, as well as Windows users can get similar, but more limited, functionality if they
the `--mask` (`-M`) argument: launch dream.py with the "winpty" program:
| Argument <img width="100" align="right"/> | Shortcut | Default | Description | ~~~
| ----------------------------------------- | ---------- | ------- | ------------------------------------------------------------------------------------------------ | > winpty python scripts\dream.py
| `--init_mask <path>` | `-M<path>` | `None` | Path to an image the same size as the initial_image, with areas for inpainting made transparent. | ~~~
## Command-line editing and completion On the Mac and Linux platforms, when you exit dream.py, the last 1000
lines of your command-line history will be saved. When you restart
dream.py, you can access the saved history using the up-arrow key.
If you are on a Macintosh or Linux machine, the command-line offers convenient In addition, limited command-line completion is installed. In various
history tracking, editing, and command completion. contexts, you can start typing your command and press tab. A list of
potential completions will be presented to you. You can then type a
little more, hit tab again, and eventually autocomplete what you want.
- To scroll through previous commands and potentially edit/reuse them, use the When specifying file paths using the one-letter shortcuts, the CLI
++up++ and ++down++ cursor keys. will attempt to complete pathnames for you. This is most handy for the
- To edit the current command, use the ++left++ and ++right++ cursor keys to -I (init image) and -M (init mask) paths. To initiate completion, start
position the cursor, and then ++backspace++, ++delete++ or ++insert++ the path with a slash ("/") or "./". For example:
characters.
- To move to the very beginning of the command, type ++ctrl+a++ (or
++command+a++ on the Mac)
- To move to the end of the command, type ++ctrl+e++.
- To cut a section of the command, position the cursor where you want to start
cutting and type ++ctrl+k++.
- To paste a cut section back in, position the cursor where you want to paste,
and type ++ctrl+y++
Windows users can get similar, but more limited, functionality if they launch ~~~
`dream.py` with the "winpty" program: dream> zebra with a mustache -I./test-pictures<TAB>
```batch
winpty python scripts\dream.py
```
On the Mac and Linux platforms, when you exit `dream.py`, the last 1000 lines of
your command-line history will be saved. When you restart `dream.py`, you can
access the saved history using the ++up++ key.
In addition, limited command-line completion is installed. In various contexts,
you can start typing your command and press tab. A list of potential completions
will be presented to you. You can then type a little more, hit tab again, and
eventually autocomplete what you want.
When specifying file paths using the one-letter shortcuts, the CLI will attempt
to complete pathnames for you. This is most handy for the `-I` (init image) and
`-M` (init mask) paths. To initiate completion, start the path with a slash `/`
or `./`, for example:
```bash
dream> "zebra with a mustache" -I./test-pictures<TAB>
-I./test-pictures/Lincoln-and-Parrot.png -I./test-pictures/zebra.jpg -I./test-pictures/madonna.png -I./test-pictures/Lincoln-and-Parrot.png -I./test-pictures/zebra.jpg -I./test-pictures/madonna.png
-I./test-pictures/bad-sketch.png -I./test-pictures/man_with_eagle/ -I./test-pictures/bad-sketch.png -I./test-pictures/man_with_eagle/
``` ```

35
docs/help/TROUBLESHOOT.md
View File

@ -15,12 +15,37 @@ incomplete installations or crashes during the install process.
During `conda env create -f environment.yaml`, conda hangs indefinitely. During `conda env create -f environment.yaml`, conda hangs indefinitely.
### **SOLUTION** If it is because of the last PIP step (usually stuck in the Git Clone step, you can check the detailed log by this method):
```bash
export PIP_LOG="/tmp/pip_log.txt"
touch ${PIP_LOG}
tail -f ${PIP_LOG} &
conda env create -f environment-mac.yaml --debug --verbose
killall tail
rm ${PIP_LOG}
```
Enter the stable-diffusion directory and completely remove the `src` directory and all its contents. **SOLUTION**
The safest way to do this is to enter the stable-diffusion directory and give the command
`git clean -f`. If this still doesn't fix the problem, try "conda clean -all" and then restart at Conda sometimes gets stuck at the last PIP step, in which several git repositories are
the `conda env create` step. cloned and built.
Enter the stable-diffusion directory and completely remove the `src`
directory and all its contents. The safest way to do this is to enter
the stable-diffusion directory and give the command `git clean -f`. If
this still doesn't fix the problem, try "conda clean -all" and then
restart at the `conda env create` step.
To further understand the problem to checking the install lot using this method:
```bash
export PIP_LOG="/tmp/pip_log.txt"
touch ${PIP_LOG}
tail -f ${PIP_LOG} &
conda env create -f environment-mac.yaml --debug --verbose
killall tail
rm ${PIP_LOG}
```
--- ---

5
docs/installation/INSTALL_MAC.md
View File

@ -95,7 +95,8 @@ While that is downloading, open a Terminal and run the following commands:
```{.bash .annotate title="local repo setup"} ```{.bash .annotate title="local repo setup"}
# clone the repo # clone the repo
git clone https://github.com/invoke-ai/InvokeAI.git git clone https://github.com/invoke-ai/InvokeAI.git
cd stable-diffusion
cd InvokeAI
# wait until the checkpoint file has downloaded, then proceed # wait until the checkpoint file has downloaded, then proceed
@ -124,7 +125,7 @@ ln -s "$PATH_TO_CKPT/sd-v1-4.ckpt" \
=== "Intel x86_64" === "Intel x86_64"
```bash ```bash
PIP_EXISTS_ACTION=w CONDA_SUBDIR=osx-x86_64 \ PIP_EXISTS_ACTION=w CONDA_SUBDIR=osx-64 \
conda env create \ conda env create \
-f environment-mac.yaml \ -f environment-mac.yaml \
&& conda activate ldm && conda activate ldm

1
environment-mac.yaml
View File

@ -30,6 +30,7 @@ dependencies:
- nomkl - nomkl
- numpy==1.23.2 - numpy==1.23.2
- omegaconf==2.1.1 - omegaconf==2.1.1
- openh264==2.3.0
- onnx==1.12.0 - onnx==1.12.0
- onnxruntime==1.12.1 - onnxruntime==1.12.1
- protobuf==3.20.1 - protobuf==3.20.1

694
frontend/dist/assets/index.66192cce.js vendored Normal file
View File

File diff suppressed because one or more lines are too long

4
frontend/src/app/App.tsx
View File

@ -10,15 +10,13 @@ import PromptInput from '../features/options/PromptInput';
import LogViewer from '../features/system/LogViewer'; import LogViewer from '../features/system/LogViewer';
import Loading from '../Loading'; import Loading from '../Loading';
import { useAppDispatch } from './store'; import { useAppDispatch } from './store';
import { requestAllImages, requestSystemConfig } from './socketio/actions'; import { requestSystemConfig } from './socketio/actions';
const App = () => { const App = () => {
const dispatch = useAppDispatch(); const dispatch = useAppDispatch();
const [isReady, setIsReady] = useState<boolean>(false); const [isReady, setIsReady] = useState<boolean>(false);
// Load images from the gallery once
useEffect(() => { useEffect(() => {
dispatch(requestAllImages());
dispatch(requestSystemConfig()); dispatch(requestSystemConfig());
setIsReady(true); setIsReady(true);
}, [dispatch]); }, [dispatch]);

7
frontend/src/app/invokeai.d.ts vendored
View File

@ -107,6 +107,7 @@ export declare type Metadata = SystemConfig & {
export declare type Image = { export declare type Image = {
uuid: string; uuid: string;
url: string; url: string;
mtime: number;
metadata: Metadata; metadata: Metadata;
}; };
@ -148,6 +149,7 @@ export declare type SystemConfigResponse = SystemConfig;
export declare type ImageResultResponse = { export declare type ImageResultResponse = {
url: string; url: string;
mtime: number;
metadata: Metadata; metadata: Metadata;
}; };
@ -157,7 +159,10 @@ export declare type ErrorResponse = {
}; };
export declare type GalleryImagesResponse = { export declare type GalleryImagesResponse = {
images: Array<{ url: string; metadata: Metadata }>; images: Array<Omit<Image, 'uuid'>>;
nextPage: number;
offset: number;
onlyNewImages: boolean;
}; };
export declare type ImageUrlAndUuidResponse = { export declare type ImageUrlAndUuidResponse = {

11
frontend/src/app/socketio/actions.ts
View File

@ -12,8 +12,11 @@ export const generateImage = createAction<undefined>('socketio/generateImage');
export const runESRGAN = createAction<InvokeAI.Image>('socketio/runESRGAN'); export const runESRGAN = createAction<InvokeAI.Image>('socketio/runESRGAN');
export const runGFPGAN = createAction<InvokeAI.Image>('socketio/runGFPGAN'); export const runGFPGAN = createAction<InvokeAI.Image>('socketio/runGFPGAN');
export const deleteImage = createAction<InvokeAI.Image>('socketio/deleteImage'); export const deleteImage = createAction<InvokeAI.Image>('socketio/deleteImage');
export const requestAllImages = createAction<undefined>( export const requestImages = createAction<undefined>(
'socketio/requestAllImages' 'socketio/requestImages'
);
export const requestNewImages = createAction<undefined>(
'socketio/requestNewImages'
); );
export const cancelProcessing = createAction<undefined>( export const cancelProcessing = createAction<undefined>(
'socketio/cancelProcessing' 'socketio/cancelProcessing'
@ -23,4 +26,6 @@ export const uploadInitialImage = createAction<File>(
); );
export const uploadMaskImage = createAction<File>('socketio/uploadMaskImage'); export const uploadMaskImage = createAction<File>('socketio/uploadMaskImage');
export const requestSystemConfig = createAction<undefined>('socketio/requestSystemConfig'); export const requestSystemConfig = createAction<undefined>(
'socketio/requestSystemConfig'
);

17
frontend/src/app/socketio/emitters.ts
View File

@ -83,8 +83,17 @@ const makeSocketIOEmitters = (
const { url, uuid } = imageToDelete; const { url, uuid } = imageToDelete;
socketio.emit('deleteImage', url, uuid); socketio.emit('deleteImage', url, uuid);
}, },
emitRequestAllImages: () => { emitRequestImages: () => {
socketio.emit('requestAllImages'); const { nextPage, offset } = getState().gallery;
socketio.emit('requestImages', nextPage, offset);
},
emitRequestNewImages: () => {
const { nextPage, offset, images } = getState().gallery;
if (images.length > 0) {
socketio.emit('requestImages', nextPage, offset, images[0].mtime);
} else {
socketio.emit('requestImages', nextPage, offset);
}
}, },
emitCancelProcessing: () => { emitCancelProcessing: () => {
socketio.emit('cancel'); socketio.emit('cancel');
@ -96,8 +105,8 @@ const makeSocketIOEmitters = (
socketio.emit('uploadMaskImage', file, file.name); socketio.emit('uploadMaskImage', file, file.name);
}, },
emitRequestSystemConfig: () => { emitRequestSystemConfig: () => {
socketio.emit('requestSystemConfig') socketio.emit('requestSystemConfig');
} },
}; };
}; };

33
frontend/src/app/socketio/listeners.ts
View File

@ -14,10 +14,10 @@ import {
} from '../../features/system/systemSlice'; } from '../../features/system/systemSlice';
import { import {
addGalleryImages,
addImage, addImage,
clearIntermediateImage, clearIntermediateImage,
removeImage, removeImage,
setGalleryImages,
setIntermediateImage, setIntermediateImage,
} from '../../features/gallery/gallerySlice'; } from '../../features/gallery/gallerySlice';
@ -25,6 +25,7 @@ import {
setInitialImagePath, setInitialImagePath,
setMaskPath, setMaskPath,
} from '../../features/options/optionsSlice'; } from '../../features/options/optionsSlice';
import { requestNewImages } from './actions';
/** /**
* Returns an object containing listener callbacks for socketio events. * Returns an object containing listener callbacks for socketio events.
@ -43,6 +44,7 @@ const makeSocketIOListeners = (
try { try {
dispatch(setIsConnected(true)); dispatch(setIsConnected(true));
dispatch(setCurrentStatus('Connected')); dispatch(setCurrentStatus('Connected'));
dispatch(requestNewImages());
} catch (e) { } catch (e) {
console.error(e); console.error(e);
} }
@ -53,7 +55,6 @@ const makeSocketIOListeners = (
onDisconnect: () => { onDisconnect: () => {
try { try {
dispatch(setIsConnected(false)); dispatch(setIsConnected(false));
dispatch(setIsProcessing(false));
dispatch(setCurrentStatus('Disconnected')); dispatch(setCurrentStatus('Disconnected'));
dispatch( dispatch(
@ -72,13 +73,14 @@ const makeSocketIOListeners = (
*/ */
onGenerationResult: (data: InvokeAI.ImageResultResponse) => { onGenerationResult: (data: InvokeAI.ImageResultResponse) => {
try { try {
const { url, metadata } = data; const { url, mtime, metadata } = data;
const newUuid = uuidv4(); const newUuid = uuidv4();
dispatch( dispatch(
addImage({ addImage({
uuid: newUuid, uuid: newUuid,
url, url,
mtime,
metadata: metadata, metadata: metadata,
}) })
); );
@ -99,11 +101,12 @@ const makeSocketIOListeners = (
onIntermediateResult: (data: InvokeAI.ImageResultResponse) => { onIntermediateResult: (data: InvokeAI.ImageResultResponse) => {
try { try {
const uuid = uuidv4(); const uuid = uuidv4();
const { url, metadata } = data; const { url, metadata, mtime } = data;
dispatch( dispatch(
setIntermediateImage({ setIntermediateImage({
uuid, uuid,
url, url,
mtime,
metadata, metadata,
}) })
); );
@ -123,12 +126,13 @@ const makeSocketIOListeners = (
*/ */
onESRGANResult: (data: InvokeAI.ImageResultResponse) => { onESRGANResult: (data: InvokeAI.ImageResultResponse) => {
try { try {
const { url, metadata } = data; const { url, metadata, mtime } = data;
dispatch( dispatch(
addImage({ addImage({
uuid: uuidv4(), uuid: uuidv4(),
url, url,
mtime,
metadata, metadata,
}) })
); );
@ -149,12 +153,13 @@ const makeSocketIOListeners = (
*/ */
onGFPGANResult: (data: InvokeAI.ImageResultResponse) => { onGFPGANResult: (data: InvokeAI.ImageResultResponse) => {
try { try {
const { url, metadata } = data; const { url, metadata, mtime } = data;
dispatch( dispatch(
addImage({ addImage({
uuid: uuidv4(), uuid: uuidv4(),
url, url,
mtime,
metadata, metadata,
}) })
); );
@ -209,16 +214,26 @@ const makeSocketIOListeners = (
* Callback to run when we receive a 'galleryImages' event. * Callback to run when we receive a 'galleryImages' event.
*/ */
onGalleryImages: (data: InvokeAI.GalleryImagesResponse) => { onGalleryImages: (data: InvokeAI.GalleryImagesResponse) => {
const { images } = data; const { images, nextPage, offset } = data;
/**
* the logic here ideally would be in the reducer but we have a side effect:
* generating a uuid. so the logic needs to be here, outside redux.
*/
// Generate a UUID for each image
const preparedImages = images.map((image): InvokeAI.Image => { const preparedImages = images.map((image): InvokeAI.Image => {
const { url, metadata } = image; const { url, metadata, mtime } = image;
return { return {
uuid: uuidv4(), uuid: uuidv4(),
url, url,
mtime,
metadata, metadata,
}; };
}); });
dispatch(setGalleryImages(preparedImages));
dispatch(addGalleryImages({ images: preparedImages, nextPage, offset }));
dispatch( dispatch(
addLogEntry({ addLogEntry({
timestamp: dateFormat(new Date(), 'isoDateTime'), timestamp: dateFormat(new Date(), 'isoDateTime'),

17
frontend/src/app/socketio/middleware.ts
View File

@ -24,7 +24,9 @@ import * as InvokeAI from '../invokeai';
export const socketioMiddleware = () => { export const socketioMiddleware = () => {
const { hostname, port } = new URL(window.location.href); const { hostname, port } = new URL(window.location.href);
const socketio = io(`http://${hostname}:9090`); const socketio = io(`http://${hostname}:9090`, {
timeout: 60000,
});
let areListenersSet = false; let areListenersSet = false;
@ -51,7 +53,8 @@ export const socketioMiddleware = () => {
emitRunESRGAN, emitRunESRGAN,
emitRunGFPGAN, emitRunGFPGAN,
emitDeleteImage, emitDeleteImage,
emitRequestAllImages, emitRequestImages,
emitRequestNewImages,
emitCancelProcessing, emitCancelProcessing,
emitUploadInitialImage, emitUploadInitialImage,
emitUploadMaskImage, emitUploadMaskImage,
@ -140,11 +143,17 @@ export const socketioMiddleware = () => {
break; break;
} }
case 'socketio/requestAllImages': { case 'socketio/requestImages': {
emitRequestAllImages(); emitRequestImages();
break; break;
} }
case 'socketio/requestNewImages': {
emitRequestNewImages();
break;
}
case 'socketio/cancelProcessing': { case 'socketio/cancelProcessing': {
emitCancelProcessing(); emitCancelProcessing();
break; break;

30
frontend/src/features/gallery/ImageGallery.tsx
View File

@ -1,5 +1,6 @@
import { Center, Flex, Text } from '@chakra-ui/react'; import { Button, Center, Flex, Text } from '@chakra-ui/react';
import { RootState } from '../../app/store'; import { requestImages } from '../../app/socketio/actions';
import { RootState, useAppDispatch } from '../../app/store';
import { useAppSelector } from '../../app/store'; import { useAppSelector } from '../../app/store';
import HoverableImage from './HoverableImage'; import HoverableImage from './HoverableImage';
@ -10,7 +11,7 @@ const ImageGallery = () => {
const { images, currentImageUuid } = useAppSelector( const { images, currentImageUuid } = useAppSelector(
(state: RootState) => state.gallery (state: RootState) => state.gallery
); );
const dispatch = useAppDispatch();
/** /**
* I don't like that this needs to rerender whenever the current image is changed. * I don't like that this needs to rerender whenever the current image is changed.
* What if we have a large number of images? I suppose pagination (planned) will * What if we have a large number of images? I suppose pagination (planned) will
@ -19,15 +20,22 @@ const ImageGallery = () => {
* TODO: Refactor if performance complaints, or after migrating to new API which supports pagination. * TODO: Refactor if performance complaints, or after migrating to new API which supports pagination.
*/ */
const handleClickLoadMore = () => {
dispatch(requestImages());
};
return images.length ? ( return images.length ? (
<Flex gap={2} wrap="wrap" pb={2}> <Flex direction={'column'} gap={2} pb={2}>
{[...images].reverse().map((image) => { <Flex gap={2} wrap="wrap">
const { uuid } = image; {images.map((image) => {
const isSelected = currentImageUuid === uuid; const { uuid } = image;
return ( const isSelected = currentImageUuid === uuid;
<HoverableImage key={uuid} image={image} isSelected={isSelected} /> return (
); <HoverableImage key={uuid} image={image} isSelected={isSelected} />
})} );
})}
</Flex>
<Button onClick={handleClickLoadMore}>Load more...</Button>
</Flex> </Flex>
) : ( ) : (
<Center height={'100%'} position={'relative'}> <Center height={'100%'} position={'relative'}>

32
frontend/src/features/gallery/gallerySlice.ts
View File

@ -8,11 +8,15 @@ export interface GalleryState {
currentImageUuid: string; currentImageUuid: string;
images: Array<InvokeAI.Image>; images: Array<InvokeAI.Image>;
intermediateImage?: InvokeAI.Image; intermediateImage?: InvokeAI.Image;
nextPage: number;
offset: number;
} }
const initialState: GalleryState = { const initialState: GalleryState = {
currentImageUuid: '', currentImageUuid: '',
images: [], images: [],
nextPage: 1,
offset: 0,
}; };
export const gallerySlice = createSlice({ export const gallerySlice = createSlice({
@ -50,7 +54,7 @@ export const gallerySlice = createSlice({
* Clamp the new index to ensure it is valid.. * Clamp the new index to ensure it is valid..
*/ */
const newCurrentImageIndex = clamp( const newCurrentImageIndex = clamp(
imageToDeleteIndex - 1, imageToDeleteIndex,
0, 0,
newImages.length - 1 newImages.length - 1
); );
@ -67,10 +71,11 @@ export const gallerySlice = createSlice({
state.images = newImages; state.images = newImages;
}, },
addImage: (state, action: PayloadAction<InvokeAI.Image>) => { addImage: (state, action: PayloadAction<InvokeAI.Image>) => {
state.images.push(action.payload); state.images.unshift(action.payload);
state.currentImageUuid = action.payload.uuid; state.currentImageUuid = action.payload.uuid;
state.intermediateImage = undefined; state.intermediateImage = undefined;
state.currentImage = action.payload; state.currentImage = action.payload;
state.offset += 1
}, },
setIntermediateImage: (state, action: PayloadAction<InvokeAI.Image>) => { setIntermediateImage: (state, action: PayloadAction<InvokeAI.Image>) => {
state.intermediateImage = action.payload; state.intermediateImage = action.payload;
@ -78,13 +83,24 @@ export const gallerySlice = createSlice({
clearIntermediateImage: (state) => { clearIntermediateImage: (state) => {
state.intermediateImage = undefined; state.intermediateImage = undefined;
}, },
setGalleryImages: (state, action: PayloadAction<Array<InvokeAI.Image>>) => { addGalleryImages: (
const newImages = action.payload; state,
if (newImages.length) { action: PayloadAction<{
const newCurrentImage = newImages[newImages.length - 1]; images: Array<InvokeAI.Image>;
state.images = newImages; nextPage: number;
offset: number;
}>
) => {
const { images, nextPage, offset } = action.payload;
if (images.length) {
const newCurrentImage = images[0];
state.images = state.images
.concat(images)
.sort((a, b) => b.mtime - a.mtime);
state.currentImage = newCurrentImage; state.currentImage = newCurrentImage;
state.currentImageUuid = newCurrentImage.uuid; state.currentImageUuid = newCurrentImage.uuid;
state.nextPage = nextPage;
state.offset = offset;
} }
}, },
}, },
@ -95,7 +111,7 @@ export const {
clearIntermediateImage, clearIntermediateImage,
removeImage, removeImage,
setCurrentImage, setCurrentImage,
setGalleryImages, addGalleryImages,
setIntermediateImage, setIntermediateImage,
} = gallerySlice.actions; } = gallerySlice.actions;

19
frontend/src/features/system/SettingsModal.tsx
View File

@ -32,8 +32,16 @@ import { cloneElement, ReactElement } from 'react';
const systemSelector = createSelector( const systemSelector = createSelector(
(state: RootState) => state.system, (state: RootState) => state.system,
(system: SystemState) => { (system: SystemState) => {
const { shouldDisplayInProgress, shouldConfirmOnDelete, shouldDisplayGuides } = system; const {
return { shouldDisplayInProgress, shouldConfirmOnDelete, shouldDisplayGuides }; shouldDisplayInProgress,
shouldConfirmOnDelete,
shouldDisplayGuides,
} = system;
return {
shouldDisplayInProgress,
shouldConfirmOnDelete,
shouldDisplayGuides,
};
}, },
{ {
memoizeOptions: { resultEqualityCheck: isEqual }, memoizeOptions: { resultEqualityCheck: isEqual },
@ -64,8 +72,11 @@ const SettingsModal = ({ children }: SettingsModalProps) => {
onClose: onRefreshModalClose, onClose: onRefreshModalClose,
} = useDisclosure(); } = useDisclosure();
const { shouldDisplayInProgress, shouldConfirmOnDelete, shouldDisplayGuides } = const {
useAppSelector(systemSelector); shouldDisplayInProgress,
shouldConfirmOnDelete,
shouldDisplayGuides,
} = useAppSelector(systemSelector);
const dispatch = useAppDispatch(); const dispatch = useAppDispatch();

15
frontend/src/features/system/systemSlice.ts
View File

@ -1,7 +1,7 @@
import { createSlice } from '@reduxjs/toolkit'; import { createSlice } from '@reduxjs/toolkit';
import type { PayloadAction } from '@reduxjs/toolkit'; import type { PayloadAction } from '@reduxjs/toolkit';
import { ExpandedIndex } from '@chakra-ui/react'; import { ExpandedIndex } from '@chakra-ui/react';
import * as InvokeAI from '../../app/invokeai' import * as InvokeAI from '../../app/invokeai';
export type LogLevel = 'info' | 'warning' | 'error'; export type LogLevel = 'info' | 'warning' | 'error';
@ -15,7 +15,9 @@ export interface Log {
[index: number]: LogEntry; [index: number]: LogEntry;
} }
export interface SystemState extends InvokeAI.SystemStatus, InvokeAI.SystemConfig { export interface SystemState
extends InvokeAI.SystemStatus,
InvokeAI.SystemConfig {
shouldDisplayInProgress: boolean; shouldDisplayInProgress: boolean;
log: Array<LogEntry>; log: Array<LogEntry>;
shouldShowLogViewer: boolean; shouldShowLogViewer: boolean;
@ -31,7 +33,6 @@ export interface SystemState extends InvokeAI.SystemStatus, InvokeAI.SystemConfi
totalIterations: number; totalIterations: number;
currentStatus: string; currentStatus: string;
currentStatusHasSteps: boolean; currentStatusHasSteps: boolean;
shouldDisplayGuides: boolean; shouldDisplayGuides: boolean;
} }
@ -51,7 +52,7 @@ const initialSystemState = {
totalSteps: 0, totalSteps: 0,
currentIteration: 0, currentIteration: 0,
totalIterations: 0, totalIterations: 0,
currentStatus: '', currentStatus: 'Disconnected',
currentStatusHasSteps: false, currentStatusHasSteps: false,
model: '', model: '',
model_id: '', model_id: '',
@ -107,6 +108,12 @@ export const systemSlice = createSlice({
}, },
setIsConnected: (state, action: PayloadAction<boolean>) => { setIsConnected: (state, action: PayloadAction<boolean>) => {
state.isConnected = action.payload; state.isConnected = action.payload;
state.isProcessing = false;
state.currentStep = 0;
state.totalSteps = 0;
state.currentIteration = 0;
state.totalIterations = 0;
state.currentStatusHasSteps = false;
}, },
setSocketId: (state, action: PayloadAction<string>) => { setSocketId: (state, action: PayloadAction<string>) => {
state.socketId = action.payload; state.socketId = action.payload;

8
ldm/dream/args.py
View File

@ -339,6 +339,12 @@ class Args(object):
action='store_true', action='store_true',
help='Deprecated way to set --precision=float32', help='Deprecated way to set --precision=float32',
) )
model_group.add_argument(
'--free_gpu_mem',
dest='free_gpu_mem',
action='store_true',
help='Force free gpu memory before final decoding',
)
model_group.add_argument( model_group.add_argument(
'--precision', '--precision',
dest='precision', dest='precision',
@ -588,7 +594,7 @@ class Args(object):
'--upscale', '--upscale',
nargs='+', nargs='+',
type=float, type=float,
help='Scale factor (2, 4) for upscaling final output followed by upscaling strength (0-1.0). If strength not specified, defaults to 0.75', help='Scale factor (1, 2, 3, 4, etc..) for upscaling final output followed by upscaling strength (0-1.0). If strength not specified, defaults to 0.75',
default=None, default=None,
) )
postprocessing_group.add_argument( postprocessing_group.add_argument(

134
ldm/dream/generator/embiggen.py
View File

@ -4,18 +4,42 @@ and generates with ldm.dream.generator.img2img
''' '''
import torch import torch
import numpy as np import numpy as np
from PIL import Image from tqdm import trange
from ldm.dream.generator.base import Generator from PIL import Image
from ldm.models.diffusion.ddim import DDIMSampler from ldm.dream.generator.base import Generator
from ldm.dream.generator.img2img import Img2Img from ldm.dream.generator.img2img import Img2Img
from ldm.dream.devices import choose_autocast
class Embiggen(Generator): class Embiggen(Generator):
def __init__(self, model, precision): def __init__(self, model, precision):
super().__init__(model, precision) super().__init__(model, precision)
self.init_latent = None self.init_latent = None
# Replace generate because Embiggen doesn't need/use most of what it does normallly
def generate(self,prompt,iterations=1,seed=None,
image_callback=None, step_callback=None,
**kwargs):
scope = choose_autocast(self.precision)
make_image = self.get_make_image(
prompt,
step_callback = step_callback,
**kwargs
)
results = []
seed = seed if seed else self.new_seed()
# Noise will be generated by the Img2Img generator when called
with scope(self.model.device.type), self.model.ema_scope():
for n in trange(iterations, desc='Generating'):
# make_image will call Img2Img which will do the equivalent of get_noise itself
image = make_image()
results.append([image, seed])
if image_callback is not None:
image_callback(image, seed)
seed = self.new_seed()
return results
@torch.no_grad() @torch.no_grad()
def get_make_image( def get_make_image(
self, self,
@ -151,8 +175,19 @@ class Embiggen(Generator):
# Clamp values to max 255 # Clamp values to max 255
if distanceToLR > 255: if distanceToLR > 255:
distanceToLR = 255 distanceToLR = 255
# Place the pixel as invert of distance #Place the pixel as invert of distance
agradientC.putpixel((x, y), int(255 - distanceToLR)) agradientC.putpixel((x, y), round(255 - distanceToLR))
# Create alternative asymmetric diagonal corner to use on "tailing" intersections to prevent hard edges
# Fits for a left-fading gradient on the bottom side and full opacity on the right side.
agradientAsymC = Image.new('L', (256, 256))
for y in range(256):
for x in range(256):
value = round(max(0, x-(255-y)) * (255 / max(1,y)))
#Clamp values
value = max(0, value)
value = min(255, value)
agradientAsymC.putpixel((x, y), value)
# Create alpha layers default fully white # Create alpha layers default fully white
alphaLayerL = Image.new("L", (width, height), 255) alphaLayerL = Image.new("L", (width, height), 255)
@ -163,8 +198,13 @@ class Embiggen(Generator):
alphaLayerT.paste(agradientT, (0, 0)) alphaLayerT.paste(agradientT, (0, 0))
alphaLayerLTC.paste(agradientL, (0, 0)) alphaLayerLTC.paste(agradientL, (0, 0))
alphaLayerLTC.paste(agradientT, (0, 0)) alphaLayerLTC.paste(agradientT, (0, 0))
alphaLayerLTC.paste(agradientC.resize( alphaLayerLTC.paste(agradientC.resize((overlap_size_x, overlap_size_y)), (0, 0))
(overlap_size_x, overlap_size_y)), (0, 0)) # make masks with an asymmetric upper-right corner so when the curved transparent corner of the next tile
# to its right is placed it doesn't reveal a hard trailing semi-transparent edge in the overlapping space
alphaLayerTaC = alphaLayerT.copy()
alphaLayerTaC.paste(agradientAsymC.rotate(270).resize((overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
alphaLayerLTaC = alphaLayerLTC.copy()
alphaLayerLTaC.paste(agradientAsymC.rotate(270).resize((overlap_size_x, overlap_size_y)), (width - overlap_size_x, 0))
if embiggen_tiles: if embiggen_tiles:
# Individual unconnected sides # Individual unconnected sides
@ -242,7 +282,7 @@ class Embiggen(Generator):
del agradientT del agradientT
del agradientC del agradientC
def make_image(x_T): def make_image():
# Make main tiles ------------------------------------------------- # Make main tiles -------------------------------------------------
if embiggen_tiles: if embiggen_tiles:
print(f'>> Making {len(embiggen_tiles)} Embiggen tiles...') print(f'>> Making {len(embiggen_tiles)} Embiggen tiles...')
@ -251,7 +291,20 @@ class Embiggen(Generator):
f'>> Making {(emb_tiles_x * emb_tiles_y)} Embiggen tiles ({emb_tiles_x}x{emb_tiles_y})...') f'>> Making {(emb_tiles_x * emb_tiles_y)} Embiggen tiles ({emb_tiles_x}x{emb_tiles_y})...')
emb_tile_store = [] emb_tile_store = []
# Although we could use the same seed for every tile for determinism, at higher strengths this may
# produce duplicated structures for each tile and make the tiling effect more obvious
# instead track and iterate a local seed we pass to Img2Img
seed = self.seed
seedintlimit = np.iinfo(np.uint32).max - 1 # only retreive this one from numpy
for tile in range(emb_tiles_x * emb_tiles_y): for tile in range(emb_tiles_x * emb_tiles_y):
# Don't iterate on first tile
if tile != 0:
if seed < seedintlimit:
seed += 1
else:
seed = 0
# Determine if this is a re-run and replace # Determine if this is a re-run and replace
if embiggen_tiles and not tile in embiggen_tiles: if embiggen_tiles and not tile in embiggen_tiles:
continue continue
@ -294,21 +347,20 @@ class Embiggen(Generator):
tile_results = gen_img2img.generate( tile_results = gen_img2img.generate(
prompt, prompt,
iterations=1, iterations = 1,
seed=self.seed, seed = seed,
sampler=sampler, sampler = sampler,
steps=steps, steps = steps,
cfg_scale=cfg_scale, cfg_scale = cfg_scale,
conditioning=conditioning, conditioning = conditioning,
ddim_eta=ddim_eta, ddim_eta = ddim_eta,
image_callback=None, # called only after the final image is generated image_callback = None, # called only after the final image is generated
step_callback=step_callback, # called after each intermediate image is generated step_callback = step_callback, # called after each intermediate image is generated
width=width, width = width,
height=height, height = height,
init_img=init_img, # img2img doesn't need this, but it might in the future init_image = newinitimage, # notice that init_image is different from init_img
init_image=newinitimage, # notice that init_image is different from init_img mask_image = None,
mask_image=None, strength = strength,
strength=strength,
) )
emb_tile_store.append(tile_results[0][0]) emb_tile_store.append(tile_results[0][0])
@ -381,24 +433,24 @@ class Embiggen(Generator):
# bottom of image # bottom of image
elif emb_row_i == emb_tiles_y - 1: elif emb_row_i == emb_tiles_y - 1:
if emb_column_i == 0: if emb_column_i == 0:
if (tile+1) in embiggen_tiles: # Look-ahead right if (tile+1) in embiggen_tiles: # Look-ahead right
intileimage.putalpha(alphaLayerT) intileimage.putalpha(alphaLayerTaC)
else: else:
intileimage.putalpha(alphaLayerRTC) intileimage.putalpha(alphaLayerRTC)
elif emb_column_i == emb_tiles_x - 1: elif emb_column_i == emb_tiles_x - 1:
# No tiles to look ahead to # No tiles to look ahead to
intileimage.putalpha(alphaLayerLTC) intileimage.putalpha(alphaLayerLTC)
else: else:
if (tile+1) in embiggen_tiles: # Look-ahead right if (tile+1) in embiggen_tiles: # Look-ahead right
intileimage.putalpha(alphaLayerLTC) intileimage.putalpha(alphaLayerLTaC)
else: else:
intileimage.putalpha(alphaLayerABB) intileimage.putalpha(alphaLayerABB)
# vertical middle of image # vertical middle of image
else: else:
if emb_column_i == 0: if emb_column_i == 0:
if (tile+1) in embiggen_tiles: # Look-ahead right if (tile+1) in embiggen_tiles: # Look-ahead right
if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down
intileimage.putalpha(alphaLayerT) intileimage.putalpha(alphaLayerTaC)
else: else:
intileimage.putalpha(alphaLayerTB) intileimage.putalpha(alphaLayerTB)
elif (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down only elif (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down only
@ -411,9 +463,9 @@ class Embiggen(Generator):
else: else:
intileimage.putalpha(alphaLayerABR) intileimage.putalpha(alphaLayerABR)
else: else:
if (tile+1) in embiggen_tiles: # Look-ahead right if (tile+1) in embiggen_tiles: # Look-ahead right
if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down if (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down
intileimage.putalpha(alphaLayerLTC) intileimage.putalpha(alphaLayerLTaC)
else: else:
intileimage.putalpha(alphaLayerABR) intileimage.putalpha(alphaLayerABR)
elif (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down only elif (tile+emb_tiles_x) in embiggen_tiles: # Look-ahead down only
@ -425,9 +477,15 @@ class Embiggen(Generator):
if emb_row_i == 0 and emb_column_i >= 1: if emb_row_i == 0 and emb_column_i >= 1:
intileimage.putalpha(alphaLayerL) intileimage.putalpha(alphaLayerL)
elif emb_row_i >= 1 and emb_column_i == 0: elif emb_row_i >= 1 and emb_column_i == 0:
intileimage.putalpha(alphaLayerT) if emb_column_i + 1 == emb_tiles_x: # If we don't have anything that can be placed to the right
intileimage.putalpha(alphaLayerT)
else:
intileimage.putalpha(alphaLayerTaC)
else: else:
intileimage.putalpha(alphaLayerLTC) if emb_column_i + 1 == emb_tiles_x: # If we don't have anything that can be placed to the right
intileimage.putalpha(alphaLayerLTC)
else:
intileimage.putalpha(alphaLayerLTaC)
# Layer tile onto final image # Layer tile onto final image
outputsuperimage.alpha_composite(intileimage, (left, top)) outputsuperimage.alpha_composite(intileimage, (left, top))
else: else:

8
ldm/dream/generator/txt2img.py
View File

@ -28,6 +28,10 @@ class Txt2Img(Generator):
height // self.downsampling_factor, height // self.downsampling_factor,
width // self.downsampling_factor, width // self.downsampling_factor,
] ]
if self.free_gpu_mem and self.model.model.device != self.model.device:
self.model.model.to(self.model.device)
samples, _ = sampler.sample( samples, _ = sampler.sample(
batch_size = 1, batch_size = 1,
S = steps, S = steps,
@ -41,6 +45,10 @@ class Txt2Img(Generator):
img_callback = step_callback, img_callback = step_callback,
threshold = threshold, threshold = threshold,
) )
if self.free_gpu_mem:
self.model.model.to("cpu")
return self.sample_to_image(samples) return self.sample_to_image(samples)
return make_image return make_image

70
ldm/dream/restoration/realesrgan.py
View File

@ -14,73 +14,53 @@ class ESRGAN():
else: else:
use_half_precision = True use_half_precision = True
def load_esrgan_bg_upsampler(self, upsampler_scale): def load_esrgan_bg_upsampler(self):
if not torch.cuda.is_available(): # CPU or MPS on M1 if not torch.cuda.is_available(): # CPU or MPS on M1
use_half_precision = False use_half_precision = False
else: else:
use_half_precision = True use_half_precision = True
model_path = { from realesrgan.archs.srvgg_arch import SRVGGNetCompact
2: 'https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth', from realesrgan import RealESRGANer
4: 'https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pth',
}
if upsampler_scale not in model_path: model = SRVGGNetCompact(num_in_ch=3, num_out_ch=3, num_feat=64, num_conv=32, upscale=4, act_type='prelu')
return None model_path = 'https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.5.0/realesr-general-x4v3.pth'
else: scale = 4
from basicsr.archs.rrdbnet_arch import RRDBNet
from realesrgan import RealESRGANer
if upsampler_scale == 4: bg_upsampler = RealESRGANer(
model = RRDBNet( scale=scale,
num_in_ch=3, model_path=model_path,
num_out_ch=3, model=model,
num_feat=64, tile=self.bg_tile_size,
num_block=23, tile_pad=10,
num_grow_ch=32, pre_pad=0,
scale=4, half=use_half_precision,
) )
if upsampler_scale == 2:
model = RRDBNet(
num_in_ch=3,
num_out_ch=3,
num_feat=64,
num_block=23,
num_grow_ch=32,
scale=2,
)
bg_upsampler = RealESRGANer(
scale=upsampler_scale,
model_path=model_path[upsampler_scale],
model=model,
tile=self.bg_tile_size,
tile_pad=10,
pre_pad=0,
half=use_half_precision,
)
return bg_upsampler return bg_upsampler
def process(self, image, strength: float, seed: str = None, upsampler_scale: int = 2): def process(self, image, strength: float, seed: str = None, upsampler_scale: int = 2):
if seed is not None:
print(
f'>> Real-ESRGAN Upscaling seed:{seed} : scale:{upsampler_scale}x'
)
with warnings.catch_warnings(): with warnings.catch_warnings():
warnings.filterwarnings('ignore', category=DeprecationWarning) warnings.filterwarnings('ignore', category=DeprecationWarning)
warnings.filterwarnings('ignore', category=UserWarning) warnings.filterwarnings('ignore', category=UserWarning)
try: try:
upsampler = self.load_esrgan_bg_upsampler(upsampler_scale) upsampler = self.load_esrgan_bg_upsampler()
except Exception: except Exception:
import traceback import traceback
import sys import sys
print('>> Error loading Real-ESRGAN:', file=sys.stderr) print('>> Error loading Real-ESRGAN:', file=sys.stderr)
print(traceback.format_exc(), file=sys.stderr) print(traceback.format_exc(), file=sys.stderr)
if upsampler_scale == 0:
print('>> Real-ESRGAN: Invalid scaling option. Image not upscaled.')
return image
if seed is not None:
print(
f'>> Real-ESRGAN Upscaling seed:{seed} : scale:{upsampler_scale}x'
)
output, _ = upsampler.enhance( output, _ = upsampler.enhance(
np.array(image, dtype=np.uint8), np.array(image, dtype=np.uint8),
outscale=upsampler_scale, outscale=upsampler_scale,

24
ldm/generate.py
View File

@ -507,11 +507,8 @@ class Generate:
prompt = None prompt = None
try: try:
args = metadata_from_png(image_path) args = metadata_from_png(image_path)
if len(args) > 1: seed = args.seed
print("* Can't postprocess a grid") prompt = args.prompt
return
seed = args[0].seed
prompt = args[0].prompt
print(f'>> retrieved seed {seed} and prompt "{prompt}" from {image_path}') print(f'>> retrieved seed {seed} and prompt "{prompt}" from {image_path}')
except: except:
m = re.search('(\d+)\.png$',image_path) m = re.search('(\d+)\.png$',image_path)
@ -665,6 +662,7 @@ class Generate:
if not self.generators.get('txt2img'): if not self.generators.get('txt2img'):
from ldm.dream.generator.txt2img import Txt2Img from ldm.dream.generator.txt2img import Txt2Img
self.generators['txt2img'] = Txt2Img(self.model, self.precision) self.generators['txt2img'] = Txt2Img(self.model, self.precision)
self.generators['txt2img'].free_gpu_mem = self.free_gpu_mem
return self.generators['txt2img'] return self.generators['txt2img']
def _make_inpaint(self): def _make_inpaint(self):
@ -733,14 +731,6 @@ class Generate:
for r in image_list: for r in image_list:
image, seed = r image, seed = r
try: try:
if upscale is not None:
if self.esrgan is not None:
if len(upscale) < 2:
upscale.append(0.75)
image = self.esrgan.process(
image, upscale[1], seed, int(upscale[0]))
else:
print(">> ESRGAN is disabled. Image not upscaled.")
if strength > 0: if strength > 0:
if self.gfpgan is not None or self.codeformer is not None: if self.gfpgan is not None or self.codeformer is not None:
if facetool == 'gfpgan': if facetool == 'gfpgan':
@ -756,6 +746,14 @@ class Generate:
image = self.codeformer.process(image=image, strength=strength, device=cf_device, seed=seed, fidelity=codeformer_fidelity) image = self.codeformer.process(image=image, strength=strength, device=cf_device, seed=seed, fidelity=codeformer_fidelity)
else: else:
print(">> Face Restoration is disabled.") print(">> Face Restoration is disabled.")
if upscale is not None:
if self.esrgan is not None:
if len(upscale) < 2:
upscale.append(0.75)
image = self.esrgan.process(
image, upscale[1], seed, int(upscale[0]))
else:
print(">> ESRGAN is disabled. Image not upscaled.")
except Exception as e: except Exception as e:
print( print(
f'>> Error running RealESRGAN or GFPGAN. Your image was not upscaled.\n{e}' f'>> Error running RealESRGAN or GFPGAN. Your image was not upscaled.\n{e}'

33
notebooks/Stable-Diffusion-local-Windows.ipynb
View File

@ -13,7 +13,7 @@
"source": [ "source": [
"Note that you will need NVIDIA drivers, Python 3.10, and Git installed\n", "Note that you will need NVIDIA drivers, Python 3.10, and Git installed\n",
"beforehand - simplified\n", "beforehand - simplified\n",
"[step-by-step instructions](https://github.com/lstein/stable-diffusion/wiki/Easy-peasy-Windows-install)\n", "[step-by-step instructions](https://github.com/invoke-ai/InvokeAI/wiki/Easy-peasy-Windows-install)\n",
"are available in the wiki (you'll only need steps 1, 2, & 3 )" "are available in the wiki (you'll only need steps 1, 2, & 3 )"
] ]
}, },
@ -40,8 +40,8 @@
"outputs": [], "outputs": [],
"source": [ "source": [
"%%cmd\n", "%%cmd\n",
"git clone https://github.com/lstein/stable-diffusion.git\n", "git clone https://github.com/invoke-ai/InvokeAI.git\n",
"cd /content/stable-diffusion/\n", "cd /content/InvokeAI/\n",
"git checkout --quiet development" "git checkout --quiet development"
] ]
}, },
@ -52,14 +52,14 @@
"outputs": [], "outputs": [],
"source": [ "source": [
"%%cmd\n", "%%cmd\n",
"pew new --python 3.10 -r requirements-lin-win-colab-CUDA.txt --dont-activate stable-diffusion" "pew new --python 3.10 -r requirements-lin-win-colab-CUDA.txt --dont-activate invoke-ai"
] ]
}, },
{ {
"cell_type": "markdown", "cell_type": "markdown",
"metadata": {}, "metadata": {},
"source": [ "source": [
"# Switch the notebook kernel to the new 'stable-diffusion' environment!\n", "# Switch the notebook kernel to the new 'invoke-ai' environment!\n",
"\n", "\n",
"## VSCode: restart VSCode and come back to this cell\n", "## VSCode: restart VSCode and come back to this cell\n",
"\n", "\n",
@ -67,7 +67,7 @@
"1. Type \"Select Interpreter\" and select \"Jupyter: Select Interpreter to Start Jupyter Server\"\n", "1. Type \"Select Interpreter\" and select \"Jupyter: Select Interpreter to Start Jupyter Server\"\n",
"1. VSCode will say that it needs to install packages. Click the \"Install\" button.\n", "1. VSCode will say that it needs to install packages. Click the \"Install\" button.\n",
"1. Once the install is finished, do 1 & 2 again\n", "1. Once the install is finished, do 1 & 2 again\n",
"1. Pick 'stable-diffusion'\n", "1. Pick 'invoke-ai'\n",
"1. Run the following cell" "1. Run the following cell"
] ]
}, },
@ -77,7 +77,7 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"%cd stable-diffusion" "%cd InvokeAI"
] ]
}, },
{ {
@ -88,7 +88,7 @@
"## Jupyter/JupyterLab\n", "## Jupyter/JupyterLab\n",
"\n", "\n",
"1. Run the cell below\n", "1. Run the cell below\n",
"1. Click on the toolbar where it says \"(ipyknel)\" ↗️. You should get a pop-up asking you to \"Select Kernel\". Pick 'stable-diffusion' from the drop-down.\n" "1. Click on the toolbar where it says \"(ipyknel)\" ↗️. You should get a pop-up asking you to \"Select Kernel\". Pick 'invoke-ai' from the drop-down.\n"
] ]
}, },
{ {
@ -106,9 +106,9 @@
"source": [ "source": [
"# DO NOT RUN THIS CELL IF YOU ARE USING VSCODE!!\n", "# DO NOT RUN THIS CELL IF YOU ARE USING VSCODE!!\n",
"%%cmd\n", "%%cmd\n",
"pew workon stable-diffusion\n", "pew workon invoke-ai\n",
"pip3 install ipykernel\n", "pip3 install ipykernel\n",
"python -m ipykernel install --name=stable-diffusion" "python -m ipykernel install --name=invoke-ai"
] ]
}, },
{ {
@ -182,15 +182,20 @@
"\n", "\n",
"Now:\n", "Now:\n",
"\n", "\n",
"1. `cd` to wherever the 'stable-diffusion' directory is\n", "1. `cd` to wherever the 'InvokeAI' directory is\n",
"1. Run `pew workon stable-diffusion`\n", "1. Run `pew workon invoke-ai`\n",
"1. Run `winpty python scripts\\dream.py`" "1. Run `winpty python scripts\\dream.py`"
] ]
},
{
"cell_type": "markdown",
"metadata": {},
"source": []
} }
], ],
"metadata": { "metadata": {
"kernelspec": { "kernelspec": {
"display_name": "Python 3.10.6 ('ldm')", "display_name": "Python 3.10.6 64-bit",
"language": "python", "language": "python",
"name": "python3" "name": "python3"
}, },
@ -208,7 +213,7 @@
}, },
"vscode": { "vscode": {
"interpreter": { "interpreter": {
"hash": "a05e4574567b7bc2c98f7f9aa579f9ea5b8739b54844ab610ac85881c4be2659" "hash": "5e164cef426134bf171f386fbddecb52046b6c1479f922ab8dfdd30df05e0e80"
} }
} }
}, },

4
scripts/dream.py
View File

@ -108,6 +108,8 @@ def main():
# preload the model # preload the model
gen.load_model() gen.load_model()
#set additional option
gen.free_gpu_mem = opt.free_gpu_mem
if not infile: if not infile:
print( print(
@ -433,7 +435,7 @@ def dream_server_loop(gen, host, port, outdir, gfpgan):
f"Point your browser at http://localhost:{port} or use the host's DNS name or IP address.") f"Point your browser at http://localhost:{port} or use the host's DNS name or IP address.")
else: else:
print(">> Default host address now 127.0.0.1 (localhost). Use --host 0.0.0.0 to bind any address.") print(">> Default host address now 127.0.0.1 (localhost). Use --host 0.0.0.0 to bind any address.")
print(f">> Point your browser at http://{host}:{port}.") print(f">> Point your browser at http://{host}:{port}")
try: try:
dream_server.serve_forever() dream_server.serve_forever()

26
scripts/preload_models.py
View File

@ -49,33 +49,13 @@ except ModuleNotFoundError:
if gfpgan: if gfpgan:
print('Loading models from RealESRGAN and facexlib') print('Loading models from RealESRGAN and facexlib')
try: try:
from basicsr.archs.rrdbnet_arch import RRDBNet from realesrgan.archs.srvgg_arch import SRVGGNetCompact
from facexlib.utils.face_restoration_helper import FaceRestoreHelper from facexlib.utils.face_restoration_helper import FaceRestoreHelper
RealESRGANer(
scale=2,
model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth',
model=RRDBNet(
num_in_ch=3,
num_out_ch=3,
num_feat=64,
num_block=23,
num_grow_ch=32,
scale=2,
),
)
RealESRGANer( RealESRGANer(
scale=4, scale=4,
model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pth', model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.5.0/realesr-general-x4v3.pth',
model=RRDBNet( model = SRVGGNetCompact(num_in_ch=3, num_out_ch=3, num_feat=64, num_conv=32, upscale=4, act_type='prelu')
num_in_ch=3,
num_out_ch=3,
num_feat=64,
num_block=23,
num_grow_ch=32,
scale=4,
),
) )
FaceRestoreHelper(1, det_model='retinaface_resnet50') FaceRestoreHelper(1, det_model='retinaface_resnet50')