v1.3.0

[AUTOMATIC1111]

Features:

• add UI to edit defaults
• token merging (via dbolya/tomesd)
• settings tab rework: add a lot of additional explanations and links
• load extensions' Git metadata in parallel to loading the main program to save a ton of time during startup
• update extensions table: show branch, show date in separate column, and show version from tags if available
• TAESD - another option for cheap live previews
• allow choosing sampler and prompts for second pass of hires fix - hidden by default, enabled in settings
• calculate hashes for Lora
• add lora hashes to infotext
• when pasting infotext, use infotext's lora hashes to find local loras for <lora:xxx:1> entries whose hashes match loras the user has
• select cross attention optimization from UI

Minor:

• bump Gradio to 3.31.0
• bump PyTorch to 2.0.1 for macOS and Linux AMD
• allow setting defaults for elements in extensions' tabs
• allow selecting file type for live previews
• show "Loading..." for extra networks when displaying for the first time
• suppress ENSD infotext for samplers that don't use it
• clientside optimizations
• add options to show/hide hidden files and dirs in extra networks, and to not list models/files in hidden directories
• allow whitespace in styles.csv
• add option to reorder tabs
• move some functionality (swap resolution and set seed to -1) to client
• option to specify editor height for img2img
• button to copy image resolution into img2img width/height sliders
• switch from pyngrok to ngrok-py
• lazy-load images in extra networks UI
• set "Navigate image viewer with gamepad" option to false by default, by request
• change upscalers to download models into user-specified directory (from commandline args) rather than the default models/<...>
• allow hiding buttons in ui-config.json

Extensions:

• add /sdapi/v1/script-info api
• use Ruff to lint Python code
• use ESlint to lint Javascript code
• add/modify CFG callbacks for Self-Attention Guidance extension
• add command and endpoint for graceful server stopping
• add some locals (prompts/seeds/etc) from processing function into the Processing class as fields
• rework quoting for infotext items that have commas in them to use JSON (should be backwards compatible except for cases where it didn't work previously)
• add /sdapi/v1/refresh-loras api checkpoint post request
• tests overhaul

Bug Fixes:

• fix an issue preventing the program from starting if the user specifies a bad Gradio theme
• fix broken prompts from file script
• fix symlink scanning for extra networks
• fix --data-dir ignored when launching via webui-user.bat COMMANDLINE_ARGS
• allow web UI to be ran fully offline
• fix inability to run with --freeze-settings
• fix inability to merge checkpoint without adding metadata
• fix extra networks' save preview image not adding infotext for jpeg/webm
• remove blinking effect from text in hires fix and scale resolution preview
• make links to http://<...>.git extensions work in the extension tab
• fix bug with webui hanging at startup due to hanging git process

---

This discussion was created from the release v1.3.0.

[NoWayG5]

after the update I can't generate on my previous setting cuz out of memory. Can someone help witch news setting I need to change?

[FinleyBuhl]

are you sure the settings are the same? The error says you don't have enough enough vram to load the model and image, as there isn't enough vram left.

[NGTAuratic]

Using doggettx optimization helped, the new sdp optimizer seems to be more memory hungry. You can change it from the optimizations tab from the settings.

[NoWayG5]

are you sure the settings are the same? The error says you don't have enough enough vram to load the model and image, as there isn't enough vram left.

yeah, i just use same settings from previous generation

[tetsuoo-online]

Using doggettx optimization helped, the new sdp optimizer seems to be more memory hungry. You can change it from the optimizations tab from the settings.

Crucial information, thank you. Since the recent updates I couldn't Hires-fix upscale anything at all, actually anything above 512x960 would fail. And I have 12GB of VRAM, this worked before (up to 1600x1600).
Anyway for now Doggettx is the answer for now. Phew...

(Edit: Actually SDP works as well, just make sure to select it manually otherwise the Automatic mode will select SDP without memory efficient attention)

[Squeezitgirdle]

Traceback

I am having the same issue. Using hires fix gets stuck for a long time after the initial run and runs through tiles (esrgan), then it just sits there. If it works, it works extremely slow. Larger images that I used to be able to create all get the same cuda error.
I've tried switching to doggetx, no success. I've added xformers and it works better, and faster than doggetx, but its still a lot slower, and still can't create images as large as it used to.

I've posted my issue here: #10822

[Lling00]

Update to report an error
Launching Web UI with arguments: --xformers --no-half --no-half-vae --api
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ E:\stable-diffusion-webui-master\launch.py:38 in │
│ │
│ 35 │
│ 36 │
│ 37 if name == "main": │
│ ❱ 38 │ main() │
│ 39 │
│ │
│ E:\stable-diffusion-webui-master\launch.py:34 in main │
│ │
│ 31 │ if args.test_server: │
│ 32 │ │ configure_for_tests() │
│ 33 │ │
│ ❱ 34 │ start() │
│ 35 │
│ 36 │
│ 37 if name == "main": │
│ │
│ E:\stable-diffusion-webui-master\modules\launch_utils.py:330 in start │
│ │
│ 327 │
│ 328 def start(): │
│ 329 │ print(f"Launching {'API server' if '--nowebui' in sys.argv else 'Web UI'} with argum │
│ ❱ 330 │ import webui │
│ 331 │ if '--nowebui' in sys.argv: │
│ 332 │ │ webui.api_only() │
│ 333 │ else: │
│ │
│ E:\stable-diffusion-webui-master\webui.py:49 in │
│ │
│ 46 │ torch.long_version = torch.version │
│ 47 │ torch.version = re.search(r'[\d.]+[\d]', torch.version).group(0) │
│ 48 │
│ ❱ 49 from modules import shared, sd_samplers, upscaler, extensions, localization, ui_tempdir, │
│ 50 import modules.codeformer_model as codeformer │
│ 51 import modules.face_restoration │
│ 52 import modules.gfpgan_model as gfpgan │
│ │
│ E:\stable-diffusion-webui-master\modules\sd_samplers.py:1 in │
│ │
│ ❱ 1 from modules import sd_samplers_compvis, sd_samplers_kdiffusion, shared │
│ 2 │
│ 3 # imports for functions that previously were here and are used by other modules │
│ 4 from modules.sd_samplers_common import samples_to_image_grid, sample_to_image # noqa: F │
│ │
│ E:\stable-diffusion-webui-master\modules\sd_samplers_compvis.py:9 in │
│ │
│ 6 import torch │
│ 7 │
│ 8 from modules.shared import state │
│ ❱ 9 from modules import sd_samplers_common, prompt_parser, shared │
│ 10 import modules.models.diffusion.uni_pc │
│ 11 │
│ 12 │
│ │
│ E:\stable-diffusion-webui-master\modules\sd_samplers_common.py:5 in │
│ │
│ 2 import numpy as np │
│ 3 import torch │
│ 4 from PIL import Image │
│ ❱ 5 from modules import devices, processing, images, sd_vae_approx, sd_samplers, sd_vae_taes │
│ 6 │
│ 7 from modules.shared import opts, state │
│ 8 import modules.shared as shared │
│ │
│ E:\stable-diffusion-webui-master\modules\processing.py:15 in │
│ │
│ 12 from skimage import exposure │
│ 13 from typing import Any, Dict, List │
│ 14 │
│ ❱ 15 import modules.sd_hijack │
│ 16 from modules import devices, prompt_parser, masking, sd_samplers, lowvram, generation_pa │
│ 17 from modules.sd_hijack import model_hijack │
│ 18 from modules.shared import opts, cmd_opts, state │
│ │
│ E:\stable-diffusion-webui-master\modules\sd_hijack.py:5 in │
│ │
│ 2 from torch.nn.functional import silu │
│ 3 from types import MethodType │
│ 4 │
│ ❱ 5 import modules.textual_inversion.textual_inversion │
│ 6 from modules import devices, sd_hijack_optimizations, shared, script_callbacks, errors │
│ 7 from modules.hypernetworks import hypernetwork │
│ 8 from modules.shared import cmd_opts │
│ │
│ E:\stable-diffusion-webui-master\modules\textual_inversion\textual_inversion.py:17 in │
│ │
│ 14 from PIL import Image, PngImagePlugin │
│ 15 from torch.utils.tensorboard import SummaryWriter │
│ 16 │
│ ❱ 17 from modules import shared, devices, sd_hijack, processing, sd_models, images, sd_sample │
│ 18 import modules.textual_inversion.dataset │
│ 19 from modules.textual_inversion.learn_schedule import LearnRateScheduler │
│ 20 │
│ │
│ E:\stable-diffusion-webui-master\modules\sd_models.py:20 in │
│ │
│ 17 from modules import paths, shared, modelloader, devices, script_callbacks, sd_vae, sd_di │
│ 18 from modules.sd_hijack_inpainting import do_inpainting_hijack │
│ 19 from modules.timer import Timer │
│ ❱ 20 import tomesd │
│ 21 │
│ 22 model_dir = "Stable-diffusion" │
│ 23 model_path = os.path.abspath(os.path.join(paths.models_path, model_dir)) │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ModuleNotFoundError: No module named 'tomesd'

[Sakura-Luna]

Install tomesd.

[Lling00]

New problem after installation
CUDA SETUP: Loading binary E:\stable-diffusion-webui-master\venv\lib\site-packages\bitsandbytes\libbitsandbytes_cudaall.dll...
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ E:\stable-diffusion-webui-master\launch.py:38 in │
│ │
│ 35 │
│ 36 │
│ 37 if name == "main": │
│ ❱ 38 │ main() │
│ 39 │
│ │
│ E:\stable-diffusion-webui-master\launch.py:34 in main │
│ │
│ 31 │ if args.test_server: │
│ 32 │ │ configure_for_tests() │
│ 33 │ │
│ ❱ 34 │ start() │
│ 35 │
│ 36 │
│ 37 if name == "main": │
│ │
│ E:\stable-diffusion-webui-master\modules\launch_utils.py:334 in start │
│ │
│ 331 │ if '--nowebui' in sys.argv: │
│ 332 │ │ webui.api_only() │
│ 333 │ else: │
│ ❱ 334 │ │ webui.webui() │
│ 335 │
│ │
│ E:\stable-diffusion-webui-master\webui.py:395 in webui │
│ │
│ 392 │ │ │ FastAPI.original_setup = FastAPI.setup │
│ 393 │ │ │ FastAPI.setup = fastapi_setup │
│ 394 │ │ │
│ ❱ 395 │ │ app, local_url, share_url = shared.demo.launch( │
│ 396 │ │ │ share=cmd_opts.share, │
│ 397 │ │ │ server_name=server_name, │
│ 398 │ │ │ server_port=cmd_opts.port, │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: Blocks.launch() got an unexpected keyword argument 'allowed_paths'

[DExUS5]

What happened to the "temp" directory? the files are no longer saved there, even when I specify the exact location.