diff --git a/.gitignore b/.gitignore
index 765a2d99..a82d44f2 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,9 +1,13 @@
.idea/
+idea/
.venv/
+venv/
.gitignore/
+gitignore/
TechnicalFiles/pip_cache/
TechnicalFiles/cache/
TechnicalFiles/temp/
TechnicalFiles/logs/
inputs/
-outputs/
\ No newline at end of file
+outputs/
+ThirdPartyRepository/
\ No newline at end of file
diff --git a/Install.sh b/Install.sh
index 70bfcb6e..3852f228 100644
--- a/Install.sh
+++ b/Install.sh
@@ -4,7 +4,7 @@ CURRENT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" &> /dev/null && pwd)"
echo "Select installation type:"
echo "1. GPU"
-echo "2. CPU"
+echo "2. CPU OR MPS"
read -n 1 -p "Enter number (1 or 2): " choice
echo ""
@@ -12,10 +12,18 @@ if [ "$choice" = "2" ]; then
INSTALL_TYPE="CPU"
export BUILD_CUDA_EXT=0
export INSTALL_KERNELS=0
+
+ if system_profiler SPDisplaysDataType | grep -q "Metal"; then
+ echo "MPS is detected. Installing MPS-specific requirements."
+ MPS_MODE=true
+ else
+ MPS_MODE=false
+ fi
else
INSTALL_TYPE="GPU"
export BUILD_CUDA_EXT=1
export INSTALL_KERNELS=1
+ MPS_MODE=false
fi
clear
@@ -44,10 +52,17 @@ ERROR_LOG="$CURRENT_DIR/TechnicalFiles/logs/installation_errors.log"
touch "$ERROR_LOG"
if [ "$INSTALL_TYPE" = "CPU" ]; then
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-CPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ if [ "$MPS_MODE" = true ]; then
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-MPS.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-MPS.txt" 2>> "$ERROR_LOG"
+ else
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ fi
else
pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements.txt" 2>> "$ERROR_LOG"
pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda.txt" 2>> "$ERROR_LOG"
diff --git a/LaunchFile/app.py b/LaunchFile/app.py
index 9b18d9ce..de8f6bb0 100644
--- a/LaunchFile/app.py
+++ b/LaunchFile/app.py
@@ -4,9 +4,13 @@
import platform
import logging
import importlib
+import torch
os.chdir(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0'
-os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "expandable_segments:True"
+if torch.backends.mps.is_available():
+ os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
+elif torch.cuda.is_available():
+ os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "expandable_segments:True"
warnings.filterwarnings("ignore")
logging.getLogger("httpx").setLevel(logging.WARNING)
cache_dir = os.path.join("TechnicalFiles/cache")
@@ -15,6 +19,12 @@
temp_dir = os.path.join("TechnicalFiles/temp")
os.makedirs(temp_dir, exist_ok=True)
os.environ["TMPDIR"] = temp_dir
+unet_path = os.path.join("inputs/image/sd_models/rembg")
+os.makedirs(unet_path, exist_ok=True)
+os.environ["U2NET_HOME"] = unet_path
+roop_path = os.path.join("ThirdPartyRepository/insightface_models")
+os.makedirs(roop_path, exist_ok=True)
+os.environ["insightface"] = roop_path
sys.modules['triton'] = None
from threading import Thread
import gradio as gr
@@ -34,7 +44,6 @@
import av
import subprocess
import json
-import torch
from einops import rearrange
import random
import tempfile
@@ -75,6 +84,7 @@
import GPUtil
except ImportError:
GPUtil = None
+ pass
WinTmp = None
if sys.platform in ['win32', 'win64']:
@@ -82,6 +92,7 @@
import WinTmp
except ImportError:
WinTmp = None
+ pass
def lazy_import(module_name, fromlist):
@@ -507,10 +518,6 @@ def parse_pdf(pdf_path):
def remove_bg(src_img_path, out_img_path):
- model_path = "inputs/image/sd_models/rembg"
- os.makedirs(model_path, exist_ok=True)
-
- os.environ["U2NET_HOME"] = model_path
with open(src_img_path, "rb") as input_file:
input_data = input_file.read()
@@ -684,8 +691,18 @@ def change_audio_format(input_audio, new_format, enable_format_changer):
def load_magicprompt_model():
model_path = os.path.join("inputs", "image", "sd_models", "MagicPrompt")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
+
if not os.path.exists(model_path):
gr.Info("Downloading MagicPrompt model...")
os.makedirs(model_path, exist_ok=True)
@@ -701,7 +718,13 @@ def load_magicprompt_model():
def generate_magicprompt(prompt, max_new_tokens):
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
tokenizer, model = load_magicprompt_model()
model.to(device)
model.eval()
@@ -728,8 +751,18 @@ def generate_magicprompt(prompt, max_new_tokens):
def load_model(model_name, model_type, n_ctx, n_batch, n_ubatch, freq_base, freq_scale):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
+
if model_name:
model_path = f"inputs/text/llm_models/{model_name}"
if model_type == "Transformers":
@@ -817,8 +850,18 @@ def load_lora_model(base_model_name, lora_model_name, model_type):
lora_model_path = f"inputs/text/llm_models/lora/{lora_model_name}"
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
+
if model_type == "Llama":
model = Llama().Llama(base_model_path, n_gpu_layers=-1 if device == "cuda" else 0,
lora_path=lora_model_path)
@@ -886,7 +929,13 @@ def load_lora_model(base_model_name, lora_model_name, model_type):
def load_moondream2_model(model_id, revision):
moondream2_model_path = os.path.join("inputs", "text", model_id)
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
try:
if not os.path.exists(moondream2_model_path):
gr.Info(f"Downloading MoonDream2 model...")
@@ -916,8 +965,18 @@ def load_moondream2_model(model_id, revision):
def load_llava_next_video_model():
model_path = os.path.join("inputs", "text", "LLaVA-NeXT-Video")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
+
try:
if not os.path.exists(model_path):
gr.Info("Downloading LLaVA-NeXT-Video model...")
@@ -992,8 +1051,14 @@ def get_existing_chats():
def transcribe_audio(audio_file_path):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- whisper_model_path = "inputs/text/whisper-medium"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ whisper_model_path = "inputs/text/whisper-large-v3-turbo"
if not os.path.exists(whisper_model_path):
gr.Info("Downloading Whisper...")
os.makedirs(whisper_model_path, exist_ok=True)
@@ -1031,7 +1096,7 @@ def load_freevc_model():
def load_whisper_model():
- whisper_model_path = "inputs/text/whisper-medium"
+ whisper_model_path = "inputs/text/whisper-large-v3-turbo"
if not os.path.exists(whisper_model_path):
gr.Info("Downloading Whisper...")
os.makedirs(whisper_model_path, exist_ok=True)
@@ -1306,7 +1371,13 @@ def image_to_base64_data_uri(image_path):
flush()
elif enable_multimodal and llm_model_name == "LLaVA-NeXT-Video":
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
if llm_model_type == "Llama":
gr.Info("LLaVA-NeXT-Video is not supported with llama model type.")
return None, None
@@ -1412,12 +1483,26 @@ def image_to_base64_data_uri(image_path):
if not speaker_wav or not language:
gr.Info("Please, select a voice and language for TTS!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
tts_model = tts_model.to(device)
if input_audio:
if not whisper_model:
whisper_model = load_whisper_model()
- device = "cuda" if torch.cuda.is_available() else "cpu"
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
whisper_model = whisper_model.to(device)
if llm_model:
context = ""
@@ -1428,7 +1513,13 @@ def image_to_base64_data_uri(image_path):
context += f"AI: {ai_text}\n"
if llm_model_type in ["Transformers", "GPTQ", "AWQ", "BNB"]:
- device = "cuda" if torch.cuda.is_available() else "cpu"
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
tokenizer.pad_token = tokenizer.eos_token
tokenizer.padding_side = "left"
@@ -1617,7 +1708,13 @@ def generate_tts_stt(text, audio, tts_settings_html, speaker_wav, language, tts_
gr.Info("Please upload source and target files for voice conversion!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
progress(0.5, desc="Performing voice conversion")
freevc_model_path = load_freevc_model()
tts = TTS().TTS(model_name=freevc_model_path, progress_bar=False).to(device)
@@ -1640,7 +1737,14 @@ def generate_tts_stt(text, audio, tts_settings_html, speaker_wav, language, tts_
if not speaker_wav or not language:
gr.Info("Please select a voice and language for TTS!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
tts_model = tts_model.to(device)
progress(0.3, desc="Generating speech")
@@ -1669,7 +1773,14 @@ def generate_tts_stt(text, audio, tts_settings_html, speaker_wav, language, tts_
if not whisper_model:
progress(0.2, desc="Loading STT model")
whisper_model = load_whisper_model()
- device = "cuda" if torch.cuda.is_available() else "cpu"
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
whisper_model = whisper_model.to(device)
progress(0.5, desc="Transcribing audio")
@@ -2123,9 +2234,20 @@ def generate_image_txt2img(prompt, negative_prompt, style_name, stable_diffusion
else:
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
+
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models",
f"{stable_diffusion_model_name}")
@@ -2651,9 +2773,19 @@ def generate_image_img2img(prompt, negative_prompt, init_image, strength, stable
else:
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models",
f"{stable_diffusion_model_name}")
@@ -2935,9 +3067,19 @@ def generate_image_depth2img(prompt, negative_prompt, init_image, seed, strength
return None, None
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models", "depth")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not os.path.exists(stable_diffusion_model_path):
gr.Info("Downloading depth2img model...")
@@ -3029,9 +3171,19 @@ def generate_image_marigold(input_image, num_inference_steps, ensemble_size):
gr.Info("Marigold Normals model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
depth_pipe = diffusers().diffusers.MarigoldDepthPipeline.from_pretrained(
depth_model_path, variant=variant, torch_dtype=torch_dtype
@@ -3092,8 +3244,17 @@ def generate_image_pix2pix(prompt, negative_prompt, init_image, seed, num_infere
gr.Info("Pix2Pix model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
pipe = StableDiffusionInstructPix2PixPipeline().StableDiffusionInstructPix2PixPipeline.from_pretrained(pix2pix_model_path, torch_dtype=torch_dtype, safety_checker=None)
pipe.to(device)
@@ -3171,8 +3332,17 @@ def generate_image_controlnet(prompt, negative_prompt, init_image, sd_version, s
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models",
f"{stable_diffusion_model_name}")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if not os.path.exists(stable_diffusion_model_path):
gr.Info(f"StableDiffusion model not found: {stable_diffusion_model_path}")
@@ -3533,8 +3703,17 @@ def generate_image_upscale_latent(prompt, image_path, upscale_factor, seed, num_
gr.Info("Please, upload an initial image!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -3656,9 +3835,19 @@ def generate_image_sdxl_refiner(prompt, init_image, output_format):
gr.Info("SDXL Refiner model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
pipe = StableDiffusionXLImg2ImgPipeline().StableDiffusionXLImg2ImgPipeline.from_pretrained(
sdxl_refiner_path, torch_dtype=torch_dtype, variant=variant, use_safetensors=True
@@ -3708,9 +3897,19 @@ def generate_image_inpaint(prompt, negative_prompt, init_image, mask_image, blur
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models", "inpaint",
f"{stable_diffusion_model_name}")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not os.path.exists(stable_diffusion_model_path):
gr.Info(f"StableDiffusion model not found: {stable_diffusion_model_path}")
@@ -3925,9 +4124,19 @@ def generate_image_outpaint(prompt, negative_prompt, init_image, stable_diffusio
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models", "inpaint",
f"{stable_diffusion_model_name}")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not os.path.exists(stable_diffusion_model_path):
gr.Info(f"StableDiffusion model not found: {stable_diffusion_model_path}")
@@ -4085,9 +4294,19 @@ def generate_image_gligen(prompt, negative_prompt, gligen_phrases, gligen_boxes,
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models",
f"{stable_diffusion_model_name}")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not os.path.exists(stable_diffusion_model_path):
gr.Info(f"StableDiffusion model not found: {stable_diffusion_model_path}")
@@ -4233,8 +4452,17 @@ def generate_image_diffedit(source_prompt, source_negative_prompt, target_prompt
sd2_1_model_path = os.path.join("inputs", "image", "sd_models", "sd2-1")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if not os.path.exists(sd2_1_model_path):
gr.Info("Downloading Stable Diffusion 2.1 model...")
@@ -4242,7 +4470,12 @@ def generate_image_diffedit(source_prompt, source_negative_prompt, target_prompt
Repo.clone_from("https://huggingface.co/stabilityai/stable-diffusion-2-1", sd2_1_model_path)
gr.Info("Stable Diffusion 2.1 model downloaded")
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
try:
pipe = StableDiffusionDiffEditPipeline().StableDiffusionDiffEditPipeline.from_pretrained(
@@ -4333,8 +4566,17 @@ def generate_image_blip_diffusion(text_prompt_input, negative_prompt, cond_image
gr.Info("BlipDiffusion model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
blip_diffusion_pipe = BlipDiffusionPipeline().BlipDiffusionPipeline.from_pretrained(
blip_diffusion_path, torch_dtype=torch_dtype
).to(device)
@@ -4382,9 +4624,19 @@ def generate_image_animatediff(prompt, negative_prompt, input_video, strength, m
stable_diffusion_model_path = os.path.join("inputs", "image", "sd_models", f"{stable_diffusion_model_name}")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not os.path.exists(stable_diffusion_model_path):
gr.Info(f"StableDiffusion model not found: {stable_diffusion_model_path}")
@@ -4687,9 +4939,19 @@ def generate_video(init_image, output_format, seed, video_settings_html, motion_
gr.Info(f"StableVideoDiffusion model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
pipe = StableVideoDiffusionPipeline().StableVideoDiffusionPipeline.from_pretrained(
pretrained_model_name_or_path=video_model_path,
torch_dtype=torch_dtype,
@@ -4738,9 +5000,19 @@ def generate_video(init_image, output_format, seed, video_settings_html, motion_
gr.Info(f"i2vgen-xl model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
pipe = I2VGenXLPipeline().I2VGenXLPipeline.from_pretrained(video_model_path, torch_dtype=torch_dtype, variant=variant).to(device)
pipe.enable_model_cpu_offload()
@@ -4791,8 +5063,17 @@ def generate_image_ldm3d(prompt, negative_prompt, seed, width, height, num_infer
gr.Info("LDM3D model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
pipe = StableDiffusionLDM3DPipeline().StableDiffusionLDM3DPipeline.from_pretrained(ldm3d_model_path, torch_dtype=torch_dtype).to(device)
if seed == "" or seed is None:
@@ -4854,8 +5135,17 @@ def generate_image_sd3_txt2img(prompt, negative_prompt, model_type, diffusers_ve
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if enable_quantize:
try:
@@ -5173,8 +5463,17 @@ def generate_image_sd3_img2img(prompt, negative_prompt, init_image, strength, di
else:
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if diffusers_version == "3-Medium":
sd3_model_path = os.path.join("inputs", "image", "sd_models", "sd3")
@@ -5322,8 +5621,17 @@ def generate_image_sd3_controlnet(prompt, negative_prompt, init_image, diffusers
gr.Info(f"SD3 ControlNet {controlnet_model} model downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
controlnet = SD3ControlNetModel().SD3ControlNetModel.from_pretrained(controlnet_path, torch_dtype=torch_dtype)
if diffusers_version == "3-Medium":
@@ -5479,8 +5787,17 @@ def generate_image_sd3_inpaint(prompt, negative_prompt, diffusers_version, init_
return None, None
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if diffusers_version == "3-Medium":
sd3_model_path = os.path.join("inputs", "image", "sd_models", "sd3")
@@ -5628,9 +5945,20 @@ def generate_image_cascade(prompt, negative_prompt, seed, stop_button, width, he
gr.Info("Stable Cascade models downloaded")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
- variant = "bf16" if torch_dtype == torch.float16 else "bf32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ variant = "bf16"
+ else:
+ torch_dtype = torch.bfloat32
+ variant = "bf32"
+
prior = StableCascadePriorPipeline().StableCascadePriorPipeline.from_pretrained(os.path.join(stable_cascade_model_path, "prior"),
variant=variant, torch_dtype=torch_dtype).to(device)
decoder = StableCascadeDecoderPipeline().StableCascadeDecoderPipeline.from_pretrained(os.path.join(stable_cascade_model_path, "decoder"),
@@ -5734,8 +6062,17 @@ def combined_callback_decoder(decoder, i, t, callback_kwargs):
def generate_image_t2i_ip_adapter(prompt, negative_prompt, ip_adapter_image, stable_diffusion_model_type, stable_diffusion_model_name, seed, num_inference_steps, guidance_scale, width, height, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -5870,11 +6207,21 @@ def generate_image_ip_adapter_faceid(prompt, negative_prompt, face_image, s_scal
ip_ckpt = os.path.join(ip_ckpt_path, "ip-adapter-faceid-plusv2_sdxl.bin")
try:
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
- app = FaceAnalysis(name="buffalo_l", providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
+
+ app = FaceAnalysis(name="buffalo_l", root=roop_path, providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
app.prepare(ctx_id=0, det_size=(640, 640))
image = cv2.imread(face_image)
@@ -5955,8 +6302,17 @@ def generate_riffusion_text2image(prompt, negative_prompt, seed, stop_button, nu
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -6083,9 +6439,19 @@ def generate_image_kandinsky_txt2img(prompt, negative_prompt, version, seed, sto
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -6250,9 +6616,19 @@ def generate_image_kandinsky_img2img(prompt, negative_prompt, init_image, versio
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -6411,9 +6787,19 @@ def generate_image_kandinsky_inpaint(prompt, negative_prompt, init_image, mask_i
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not init_image or not mask_image:
gr.Info("Please upload an initial image and provide a mask image!")
@@ -6499,8 +6885,17 @@ def generate_image_flux_txt2img(prompt, model_name, quantize_model_name, enable_
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -6677,8 +7072,17 @@ def generate_image_flux_img2img(prompt, init_image, model_name, quantize_model_n
gr.Info("Please upload an initial image!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -6841,8 +7245,17 @@ def generate_image_flux_inpaint(prompt, init_image, mask_image, model_name, seed
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if not init_image or not mask_image:
gr.Info("Please upload an initial image and provide a mask image!")
@@ -6933,8 +7346,17 @@ def generate_image_flux_controlnet(prompt, init_image, base_model_name, seed, st
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if not init_image:
gr.Info("Please upload an initial image!")
@@ -7036,8 +7458,17 @@ def generate_image_hunyuandit_txt2img(prompt, negative_prompt, seed, stop_button
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7115,8 +7546,17 @@ def generate_image_hunyuandit_controlnet(prompt, negative_prompt, init_image, co
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if not init_image:
gr.Info("Please upload an initial image!")
@@ -7195,8 +7635,17 @@ def combined_callback(pipe, i, t, callback_kwargs):
def generate_image_lumina(prompt, negative_prompt, seed, num_inference_steps, guidance_scale, height, width, max_sequence_length, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7262,9 +7711,19 @@ def generate_image_lumina(prompt, negative_prompt, seed, num_inference_steps, gu
def generate_image_kolors_txt2img(prompt, negative_prompt, seed, lora_model_names, lora_scales, guidance_scale, num_inference_steps, max_sequence_length, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7355,9 +7814,19 @@ def generate_image_kolors_txt2img(prompt, negative_prompt, seed, lora_model_name
def generate_image_kolors_img2img(prompt, negative_prompt, init_image, seed, guidance_scale, num_inference_steps, max_sequence_length, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not init_image:
gr.Info("Please upload an initial image!")
@@ -7415,9 +7884,19 @@ def generate_image_kolors_img2img(prompt, negative_prompt, init_image, seed, gui
def generate_image_kolors_ip_adapter_plus(prompt, negative_prompt, ip_adapter_image, seed, guidance_scale, num_inference_steps, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if not ip_adapter_image:
gr.Info("Please upload an initial image!")
@@ -7496,8 +7975,17 @@ def generate_image_kolors_ip_adapter_plus(prompt, negative_prompt, ip_adapter_im
def generate_image_auraflow(prompt, negative_prompt, seed, lora_model_names, lora_scales, num_inference_steps, guidance_scale, height, width, max_sequence_length, enable_aurasr, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7610,8 +8098,17 @@ def generate_image_wurstchen(prompt, negative_prompt, seed, stop_button, width,
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7726,9 +8223,19 @@ def generate_image_deepfloyd_txt2img(prompt, negative_prompt, seed, stop_button,
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -7892,9 +8399,19 @@ def generate_image_deepfloyd_img2img(prompt, negative_prompt, init_image, seed,
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8055,9 +8572,19 @@ def generate_image_deepfloyd_inpaint(prompt, negative_prompt, init_image, mask_i
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8221,8 +8748,17 @@ def generate_image_pixart(prompt, negative_prompt, version, seed, stop_button, n
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8333,8 +8869,17 @@ def generate_image_cogview3plus(prompt, negative_prompt, seed, stop_button, heig
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8415,9 +8960,19 @@ def combined_callback(pipe, i, t, callback_kwargs):
def generate_image_playgroundv2(prompt, negative_prompt, seed, height, width, num_inference_steps, guidance_scale, output_format):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8594,9 +9149,19 @@ def generate_video_modelscope(prompt, negative_prompt, seed, stop_button, num_in
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8679,8 +9244,17 @@ def generate_video_zeroscope2(prompt, video_to_enhance, seed, stop_button, stren
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8838,8 +9412,17 @@ def generate_video_cogvideox_text2video(prompt, negative_prompt, cogvideox_versi
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -8923,8 +9506,17 @@ def generate_video_cogvideox_image2video(prompt, negative_prompt, init_image, se
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if not init_image:
gr.Info("Please, upload an initial image!")
@@ -8990,8 +9582,17 @@ def generate_video_cogvideox_video2video(prompt, negative_prompt, init_video, co
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.bfloat16 if device == "cuda" else torch.bfloat32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.bfloat16
+ else:
+ torch_dtype = torch.bfloat32
if not init_video:
gr.Info("Please, upload an initial video!")
@@ -9067,8 +9668,17 @@ def generate_video_latte(prompt, negative_prompt, seed, stop_button, num_inferen
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -9178,9 +9788,19 @@ def generate_3d_stablefast3d(image, texture_resolution, foreground_ratio, remesh
def generate_3d_shap_e(prompt, init_image, seed, num_inference_steps, guidance_scale, frame_size, progress=gr.Progress()):
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
- variant = "fp16" if torch_dtype == torch.float16 else "fp32"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ variant = "fp16"
+ else:
+ torch_dtype = torch.float32
+ variant = "fp32"
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -9272,8 +9892,17 @@ def generate_3d_zero123plus(input_image, num_inference_steps, output_format, pro
gr.Info("Please upload an input image!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
progress(0.1, desc="Initializing Zero123Plus")
zero123plus_model_path = os.path.join("inputs", "3D", "zero123plus")
@@ -9326,8 +9955,17 @@ def generate_stableaudio(prompt, negative_prompt, seed, stop_button, num_inferen
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -9559,8 +10197,17 @@ def generate_audio_audioldm2(prompt, negative_prompt, model_name, seed, stop_but
stop_signal = False
stop_idx = None
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
if seed == "" or seed is None:
seed = random.randint(0, 2 ** 32 - 1)
@@ -9661,8 +10308,18 @@ def generate_bark_audio(text, voice_preset, max_length, fine_temperature, coarse
try:
progress(0.3, desc="Setting up device")
- device = "cuda" if torch.cuda.is_available() else "cpu"
- torch_dtype = torch.float16 if device == "cuda" else torch.float32
+
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
+
+ if device == "cuda":
+ torch_dtype = torch.float16
+ else:
+ torch_dtype = torch.float32
progress(0.4, desc="Loading Bark model")
processor = AutoProcessor().AutoProcessor.from_pretrained(bark_model_path)
@@ -9732,7 +10389,12 @@ def process_rvc(input_audio, model_folder, f0method, f0up_key, index_rate, filte
gr.Info("Please upload an audio file!")
return None, None
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
model_path = os.path.join("inputs", "audio", "rvc_models", model_folder)
@@ -10019,7 +10681,12 @@ def generate_video_extras(input_video, enable_downscale, downscale_factor, enabl
def generate_audio_extras(input_audio, enable_format_changer, new_format, enable_audiosr, steps, guidance_scale, enable_downscale, downscale_factor, progress=gr.Progress()):
- device = "cuda" if torch.cuda.is_available() else "cpu"
+ if torch.backends.mps.is_available():
+ device = "mps"
+ elif torch.cuda.is_available():
+ device = "cuda"
+ else:
+ device = "cpu"
progress(0.1, desc="Initializing audio processing")
if not input_audio:
@@ -10561,85 +11228,130 @@ def open_outputs_folder():
os.system(f'open "{outputs_folder}"' if os.name == "darwin" else f'xdg-open "{outputs_folder}"')
-llm_models_list = [None, "Moondream2-Image", "LLaVA-NeXT-Video", "Qwen2-Audio"] + [model for model in os.listdir("inputs/text/llm_models") if not model.endswith(".txt") and model != "vikhyatk" and model != "lora"]
-llm_lora_models_list = [None] + [model for model in os.listdir("inputs/text/llm_models/lora") if not model.endswith(".txt")]
-speaker_wavs_list = [None] + [wav for wav in os.listdir("inputs/audio/voices") if not wav.endswith(".txt")]
-stable_diffusion_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models")
- if (model.endswith(".safetensors") or model.endswith(".ckpt") or model.endswith(".gguf") or not model.endswith(".txt") and not model.endswith(".py") and not os.path.isdir(os.path.join("inputs/image/sd_models")))]
+def get_llm_models():
+ return [None, "Moondream2-Image", "LLaVA-NeXT-Video", "Qwen2-Audio"] + [
+ model for model in os.listdir("inputs/text/llm_models")
+ if not model.endswith(".txt") and model != "vikhyatk" and model != "lora" and model != "avatars"
+ ]
+
+
+def get_llm_lora_models():
+ return [None] + [
+ model for model in os.listdir("inputs/text/llm_models/lora")
+ if not model.endswith(".txt")
+ ]
+
+
+def get_speaker_wavs():
+ return [None] + [
+ wav for wav in os.listdir("inputs/audio/voices")
+ if not wav.endswith(".txt")
+ ]
+
+
+def get_stable_diffusion_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/sd_models")
+ if model.endswith((".safetensors", ".ckpt", ".gguf"))
+ or (not model.endswith((".txt", ".py")) and not os.path.isdir(os.path.join("inputs/image/sd_models", model)))
+ ]
+
+
audiocraft_models_list = [None] + ["musicgen-stereo-medium", "audiogen-medium", "musicgen-stereo-melody", "musicgen-medium", "musicgen-melody", "musicgen-large",
"hybrid-magnet-medium", "magnet-medium-30sec", "magnet-medium-10sec", "audio-magnet-medium"]
-vae_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/vae") if
- model.endswith(".safetensors") or not model.endswith(".txt")]
-flux_vae_models_list = [None] + [model for model in os.listdir("inputs/image/flux/flux-vae") if
- model.endswith(".safetensors") or not model.endswith(".txt")]
-lora_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/lora") if
- model.endswith(".safetensors") or model.endswith(".pt")]
-quantized_flux_models_list = [None] + [model for model in os.listdir("inputs/image/flux/quantize-flux") if
- model.endswith(".gguf") or model.endswith(".safetensors") or not model.endswith(".txt") and not model.endswith(".safetensors") and not model.endswith(".py")]
-flux_lora_models_list = [None] + [model for model in os.listdir("inputs/image/flux/flux-lora") if
- model.endswith(".safetensors")]
-auraflow_lora_models_list = [None] + [model for model in os.listdir("inputs/image/auraflow-lora") if
- model.endswith(".safetensors")]
-kolors_lora_models_list = [None] + [model for model in os.listdir("inputs/image/kolors-lora") if
- model.endswith(".safetensors")]
-textual_inversion_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/embedding") if model.endswith(".pt") or model.endswith(".safetensors")]
-inpaint_models_list = [None] + [model for model in
- os.listdir("inputs/image/sd_models/inpaint")
- if (model.endswith(".safetensors") or model.endswith(".ckpt") or not model.endswith(".txt"))]
+
+
+def get_vae_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/sd_models/vae")
+ if model.endswith(".safetensors") or not model.endswith(".txt")
+ ]
+
+
+def get_flux_vae_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/flux/flux-vae")
+ if model.endswith(".safetensors") or not model.endswith(".txt")
+ ]
+
+
+def get_lora_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/sd_models/lora")
+ if model.endswith((".safetensors", ".pt"))
+ ]
+
+
+def get_quantized_flux_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/flux/quantize-flux")
+ if model.endswith((".gguf", ".safetensors"))
+ or (not model.endswith((".txt", ".safetensors", ".py")))
+ ]
+
+
+def get_flux_lora_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/flux/flux-lora")
+ if model.endswith(".safetensors")
+ ]
+
+
+def get_auraflow_lora_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/auraflow-lora")
+ if model.endswith(".safetensors")
+ ]
+
+
+def get_kolors_lora_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/kolors-lora")
+ if model.endswith(".safetensors")
+ ]
+
+
+def get_textual_inversion_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/sd_models/embedding")
+ if model.endswith((".pt", ".safetensors"))
+ ]
+
+
+def get_inpaint_models():
+ return [None] + [
+ model for model in os.listdir("inputs/image/sd_models/inpaint")
+ if model.endswith((".safetensors", ".ckpt")) or not model.endswith(".txt")
+ ]
+
+
controlnet_models_list = [None, "openpose", "depth", "canny", "lineart", "scribble"]
-rvc_models_list = [model_folder for model_folder in os.listdir("inputs/audio/rvc_models")
- if os.path.isdir(os.path.join("inputs/audio/rvc_models", model_folder))
- and any(file.endswith('.pth') for file in os.listdir(os.path.join("inputs/audio/rvc_models", model_folder)))]
-
-
-def reload_model_lists():
- global llm_models_list, llm_lora_models_list, speaker_wavs_list, stable_diffusion_models_list, vae_models_list, lora_models_list, quantized_flux_models_list, flux_lora_models_list, auraflow_lora_models_list, kolors_lora_models_list, textual_inversion_models_list, inpaint_models_list, rvc_models_list
-
- llm_models_list = [None, "Moondream2-Image", "LLaVA-NeXT-Video", "Qwen2-Audio"] + [model for model in os.listdir("inputs/text/llm_models") if
- not model.endswith(".txt") and model != "vikhyatk" and model != "lora"]
- llm_lora_models_list = [None] + [model for model in os.listdir("inputs/text/llm_models/lora") if
- not model.endswith(".txt")]
- speaker_wavs_list = [None] + [wav for wav in os.listdir("inputs/audio/voices") if not wav.endswith(".txt")]
- stable_diffusion_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models")
- if (model.endswith(".safetensors") or model.endswith(
- ".ckpt") or model.endswith(".gguf") or not model.endswith(".txt") and not model.endswith(
- ".py") and not os.path.isdir(os.path.join("inputs/image/sd_models")))]
- vae_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/vae") if
- model.endswith(".safetensors") or not model.endswith(".txt")]
- flux_vae_models_list = [None] + [model for model in os.listdir("inputs/image/flux/flux-vae") if
- model.endswith(".safetensors") or not model.endswith(".txt")]
- lora_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/lora") if
- model.endswith(".safetensors") or model.endswith(".pt")]
- quantized_flux_models_list = [None] + [model for model in os.listdir("inputs/image/flux/quantize-flux") if
- model.endswith(".gguf") or model.endswith(".safetensors") or not model.endswith(".txt") and not model.endswith(
- ".safetensors") and not model.endswith(".py")]
- flux_lora_models_list = [None] + [model for model in os.listdir("inputs/image/flux/flux-lora") if
- model.endswith(".safetensors")]
- auraflow_lora_models_list = [None] + [model for model in os.listdir("inputs/image/auraflow-lora") if
- model.endswith(".safetensors")]
- kolors_lora_models_list = [None] + [model for model in os.listdir("inputs/image/kolors-lora") if
- model.endswith(".safetensors")]
- textual_inversion_models_list = [None] + [model for model in os.listdir("inputs/image/sd_models/embedding") if
- model.endswith(".pt") or model.endswith(".safetensors")]
- inpaint_models_list = [None] + [model for model in
- os.listdir("inputs/image/sd_models/inpaint")
- if (model.endswith(".safetensors") or model.endswith(".ckpt") or not model.endswith(
- ".txt"))]
- rvc_models_list = [model_folder for model_folder in os.listdir("inputs/audio/rvc_models")
- if os.path.isdir(os.path.join("inputs/audio/rvc_models", model_folder))
- and any(
- file.endswith('.pth') for file in os.listdir(os.path.join("inputs/audio/rvc_models", model_folder)))]
-
- chat_files = get_existing_chats()
-
- gallery_files = get_output_files()
-
- return [llm_models_list, llm_lora_models_list, speaker_wavs_list, stable_diffusion_models_list, vae_models_list, flux_vae_models_list, lora_models_list, quantized_flux_models_list, flux_lora_models_list, auraflow_lora_models_list, kolors_lora_models_list, textual_inversion_models_list, inpaint_models_list, rvc_models_list, chat_files, gallery_files]
-
-
-def reload_interface():
- updated_lists = reload_model_lists()[:16]
- return [gr.Dropdown(choices=list) for list in updated_lists]
+
+
+def get_rvc_models():
+ return [
+ model_folder for model_folder in os.listdir("inputs/audio/rvc_models")
+ if os.path.isdir(os.path.join("inputs/audio/rvc_models", model_folder))
+ and any(file.endswith('.pth') for file in os.listdir(os.path.join("inputs/audio/rvc_models", model_folder)))
+ ]
+
+
+model_lists = {
+ "stable_diffusion": get_stable_diffusion_models(),
+ "llm": get_llm_models(),
+ "llm_lora": get_llm_lora_models(),
+ "speaker_wavs": get_speaker_wavs(),
+ "vae": get_vae_models(),
+ "flux_vae": get_flux_vae_models(),
+ "lora": get_lora_models(),
+ "quantized_flux": get_quantized_flux_models(),
+ "flux_lora": get_flux_lora_models(),
+ "auraflow_lora": get_auraflow_lora_models(),
+ "kolors_lora": get_kolors_lora_models(),
+ "textual_inversion": get_textual_inversion_models(),
+ "inpaint": get_inpaint_models(),
+ "rvc": get_rvc_models(),
+}
def create_footer():
@@ -10650,7 +11362,7 @@ def create_footer():
🦙 llama-cpp-python: 0.3.1
🖼️ stable-diffusion-cpp-python: 0.2.1
🎵 rvc-python: 0.1.5
- ℹ️ gradio: 5.4.0
+ ℹ️ gradio: 5.5.0
"""
return gr.Markdown(footer_html)
@@ -10668,8 +11380,8 @@ def create_footer():
gr.Textbox(label=_("Enter your system prompt", lang)),
gr.Audio(type="filepath", label=_("Record your request (optional)", lang)),
gr.Radio(choices=["Transformers", "GPTQ", "AWQ", "BNB", "Llama", "ExLlamaV2"], label=_("Select model type", lang), value="Transformers"),
- gr.Dropdown(choices=llm_models_list, label=_("Select LLM model", lang), value=None),
- gr.Dropdown(choices=llm_lora_models_list, label=_("Select LoRA model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["llm"], label=_("Select LLM model", lang), value=None),
+ gr.Dropdown(choices=model_lists["llm_lora"], label=_("Select LoRA model (optional)", lang), value=None),
gr.Dropdown(choices=get_existing_chats(), label=_("Select existing chat (optional)", lang), value=None)
],
additional_inputs=[
@@ -10710,7 +11422,7 @@ def create_footer():
gr.Slider(minimum=1, maximum=10, value=1, step=1, label=_("Num return sequences", lang)),
gr.Radio(choices=["txt", "json"], label=_("Select chat history format", lang), value="txt", interactive=True),
gr.HTML(_("TTS Settings
", lang)),
- gr.Dropdown(choices=speaker_wavs_list, label=_("Select voice", lang), interactive=True),
+ gr.Dropdown(choices=model_lists["speaker_wavs"], label=_("Select voice", lang), interactive=True),
gr.Dropdown(choices=["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"], label=_("Select language", lang), interactive=True),
gr.Slider(minimum=0.1, maximum=1.9, value=1.0, step=0.1, label=_("TTS Temperature", lang), interactive=True),
gr.Slider(minimum=0.01, maximum=1.0, value=0.9, step=0.01, label=_("TTS Top P", lang), interactive=True),
@@ -10742,7 +11454,7 @@ def create_footer():
],
additional_inputs=[
gr.HTML(_("TTS Settings
", lang)),
- gr.Dropdown(choices=speaker_wavs_list, label=_("Select voice", lang), interactive=True),
+ gr.Dropdown(choices=model_lists["speaker_wavs"], label=_("Select voice", lang), interactive=True),
gr.Dropdown(choices=["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"], label=_("Select language", lang), interactive=True),
gr.Slider(minimum=0.1, maximum=1.9, value=1.0, step=0.1, label=_("TTS Temperature", lang), interactive=True),
gr.Slider(minimum=0.01, maximum=1.0, value=0.9, step=0.01, label=_("TTS Top P", lang), interactive=True),
@@ -10875,12 +11587,12 @@ def create_footer():
gr.Textbox(label=_("Enter your prompt", lang), placeholder=_("+ and - for Weighting; ('p', 'p').blend(0.x, 0.x) for Blending; ['p', 'p', 'p'].and() for Conjunction", lang)),
gr.Textbox(label=_("Enter your negative prompt", lang), value=""),
gr.Dropdown(choices=list(styles.keys()), label=_("Select Style (optional)", lang), value=None),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
- gr.Dropdown(choices=vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
- gr.Dropdown(choices=lora_models_list, label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["vae"], label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["lora"], label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
- gr.Dropdown(choices=textual_inversion_models_list, label=_("Select Embedding models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["textual_inversion"], label=_("Select Embedding models (optional)", lang), value=None, multiselect=True),
gr.HTML(_("StableDiffusion Settings
", lang)),
gr.Radio(choices=["SD", "SD2", "SDXL"], label=_("Select model type", lang), value="SD"),
gr.Dropdown(choices=[
@@ -10947,12 +11659,12 @@ def create_footer():
gr.Image(label=_("Initial image", lang), type="filepath"),
gr.Slider(minimum=0.0, maximum=1.0, value=0.5, step=0.01, label=_("Strength (Initial image)", lang)),
gr.Radio(choices=["SD", "SD2", "SDXL"], label=_("Select model type", lang), value="SD"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
- gr.Dropdown(choices=vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
- gr.Dropdown(choices=lora_models_list, label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["vae"], label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["lora"], label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
- gr.Dropdown(choices=textual_inversion_models_list, label=_("Select Embedding models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["textual_inversion"], label=_("Select Embedding models (optional)", lang), value=None, multiselect=True),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
],
@@ -11074,7 +11786,7 @@ def create_footer():
"UniPCMultistepScheduler", "LCMScheduler", "DPMSolverSDEScheduler",
"TCDScheduler", "DDIMScheduler", "PNDMScheduler", "DDPMScheduler"
], label=_("Select scheduler", lang), value="EulerDiscreteScheduler"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Dropdown(choices=controlnet_models_list, label=_("Select ControlNet model", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
@@ -11165,8 +11877,8 @@ def create_footer():
"UniPCMultistepScheduler", "LCMScheduler", "DPMSolverSDEScheduler",
"TCDScheduler", "DDIMScheduler", "PNDMScheduler", "DDPMScheduler"
], label=_("Select scheduler", lang), value="EulerDiscreteScheduler"),
- gr.Dropdown(choices=inpaint_models_list, label=_("Select Inpaint model", lang), value=None),
- gr.Dropdown(choices=vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["inpaint"], label=_("Select Inpaint model", lang), value=None),
+ gr.Dropdown(choices=model_lists["vae"], label=_("Select VAE model (optional)", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
],
@@ -11199,7 +11911,7 @@ def create_footer():
gr.Textbox(label=_("Enter your negative prompt", lang), value=""),
gr.Image(label=_("Initial image", lang), type="filepath"),
gr.Radio(choices=["SD", "SD2", "SDXL"], label=_("Select model type", lang), value="SD"),
- gr.Dropdown(choices=inpaint_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["inpaint"], label=_("Select StableDiffusion model", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
@@ -11232,7 +11944,7 @@ def create_footer():
gr.Textbox(label=_("Enter GLIGEN phrases", lang), value=""),
gr.Textbox(label=_("Enter GLIGEN boxes", lang), value=""),
gr.Radio(choices=["SD", "SD2", "SDXL"], label=_("Select model type", lang), value="SD"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
@@ -11324,7 +12036,7 @@ def create_footer():
gr.Image(label=_("Initial GIF", lang), type="filepath"),
gr.Slider(minimum=0.0, maximum=1.0, value=0.5, step=0.01, label=_("Strength (Initial GIF)", lang)),
gr.Radio(choices=["sd", "sdxl"], label=_("Select model type", lang), value="sd"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
@@ -11454,15 +12166,15 @@ def create_footer():
gr.Textbox(label=_("Enter your negative prompt", lang), placeholder=_("(prompt:x.x) for Weighting", lang), value=""),
gr.Radio(choices=["Diffusers", "Safetensors"], label=_("Select model type", lang), value="Diffusers"),
gr.Radio(choices=["3-Medium", "3.5-Large", "3.5-Large-Turbo"], label=_("Select Diffusers model", lang), value="3-Medium"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
gr.Radio(choices=["SD3", "SD3.5"], label=_("Select quantize model type", lang), value="SD3"),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
],
additional_inputs=[
- gr.Dropdown(choices=vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
- gr.Dropdown(choices=lora_models_list, label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["vae"], label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["lora"], label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
gr.Dropdown(choices=["FlowMatchEulerDiscreteScheduler", "FlowMatchHeunDiscreteScheduler"], label=_("Select scheduler", lang), value="FlowMatchEulerDiscreteScheduler"),
gr.Slider(minimum=1, maximum=100, value=40, step=1, label=_("Steps", lang)),
@@ -11498,7 +12210,7 @@ def create_footer():
gr.Image(label=_("Initial image", lang), type="filepath"),
gr.Slider(minimum=0.0, maximum=1.0, value=0.8, step=0.01, label=_("Strength (Initial image)", lang)),
gr.Radio(choices=["3-Medium", "3.5-Large", "3.5-Large-Turbo"], label=_("Select Diffusers model", lang), value="3-Medium"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
gr.Radio(choices=["SD3", "SD3.5"], label=_("Select quantize model type", lang), value="SD3"),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
@@ -11649,7 +12361,7 @@ def create_footer():
gr.Textbox(label=_("Enter your negative prompt", lang), value=""),
gr.Image(label=_("IP-Adapter Image", lang), type="filepath"),
gr.Radio(choices=["SD", "SDXL"], label=_("Select model type", lang), value="SD"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
@@ -11682,7 +12394,7 @@ def create_footer():
gr.Image(label=_("Face image", lang), type="filepath"),
gr.Slider(minimum=0.1, maximum=2, value=1, step=0.1, label=_("Scale (Face image)", lang)),
gr.Radio(choices=["SD", "SDXL"], label=_("Select model type", lang), value="SD"),
- gr.Dropdown(choices=stable_diffusion_models_list, label=_("Select StableDiffusion model", lang), value=None),
+ gr.Dropdown(choices=model_lists["stable_diffusion"], label=_("Select StableDiffusion model", lang), value=None),
],
additional_inputs=[
gr.Slider(minimum=1, maximum=100, value=30, step=1, label=_("Steps", lang)),
@@ -11882,14 +12594,14 @@ def create_footer():
inputs=[
gr.Textbox(label=_("Enter your prompt", lang)),
gr.Radio(choices=["FLUX.1-schnell", "FLUX.1-dev"], label=_("Select model type", lang), value="FLUX.1-schnell"),
- gr.Dropdown(choices=quantized_flux_models_list, label=_("Select safetensors Flux model (GGUF if enabled quantize)", lang), value=None),
+ gr.Dropdown(choices=model_lists["quantized_flux"], label=_("Select safetensors Flux model (GGUF if enabled quantize)", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
],
additional_inputs=[
- gr.Dropdown(choices=flux_vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
- gr.Dropdown(choices=flux_lora_models_list, label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
+ gr.Dropdown(choices=model_lists["flux_vae"], label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["flux_lora"], label=_("Select LORA models (optional)", lang), value=None, multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
gr.Slider(minimum=0.0, maximum=10.0, value=0.0, step=0.1, label=_("Guidance Scale", lang)),
gr.Slider(minimum=256, maximum=2048, value=768, step=64, label=_("Height", lang)),
@@ -11921,15 +12633,15 @@ def create_footer():
gr.Image(label=_("Initial image", lang), type="filepath"),
gr.Dropdown(choices=["FLUX.1-schnell", "FLUX.1-dev"], label=_("Select Flux model", lang),
value="FLUX.1-schnell"),
- gr.Dropdown(choices=quantized_flux_models_list,
+ gr.Dropdown(choices=model_lists["quantized_flux"],
label=_("Select quantized Flux model (optional if enabled quantize)", lang), value=None),
gr.Checkbox(label=_("Enable Quantize", lang), value=False),
gr.Textbox(label=_("Seed (optional)", lang), value=""),
gr.Button(value=_("Stop generation", lang), interactive=True, variant="stop")
],
additional_inputs=[
- gr.Dropdown(choices=vae_models_list, label=_("Select VAE model (optional)", lang), value=None),
- gr.Dropdown(choices=flux_lora_models_list, label=_("Select LORA models (optional)", lang), value=None,
+ gr.Dropdown(choices=model_lists["flux_vae"], label=_("Select VAE model (optional)", lang), value=None),
+ gr.Dropdown(choices=model_lists["flux_lora"], label=_("Select LORA models (optional)", lang), value=None,
multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
gr.Slider(minimum=1, maximum=100, value=4, step=1, label=_("Steps", lang)),
@@ -12133,7 +12845,7 @@ def create_footer():
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
- gr.Dropdown(choices=kolors_lora_models_list, label=_("Select LORA models (optional)", lang), value=None,
+ gr.Dropdown(choices=model_lists["kolors_lora"], label=_("Select LORA models (optional)", lang), value=None,
multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
gr.Slider(minimum=1.0, maximum=20.0, value=6.5, step=0.1, label=_("Guidance Scale", lang)),
@@ -12226,7 +12938,7 @@ def create_footer():
gr.Textbox(label=_("Seed (optional)", lang), value="")
],
additional_inputs=[
- gr.Dropdown(choices=auraflow_lora_models_list, label=_("Select LORA models (optional)", lang), value=None,
+ gr.Dropdown(choices=model_lists["auraflow_lora"], label=_("Select LORA models (optional)", lang), value=None,
multiselect=True),
gr.Textbox(label=_("LoRA Scales", lang)),
gr.Slider(minimum=1, maximum=100, value=25, step=1, label=_("Steps", lang)),
@@ -12932,7 +13644,7 @@ def create_footer():
fn=process_rvc,
inputs=[
gr.Audio(label=_("Input audio", lang), type="filepath"),
- gr.Dropdown(choices=rvc_models_list, label=_("Select RVC model", lang), value=None)
+ gr.Dropdown(choices=model_lists["rvc"], label=_("Select RVC model", lang), value=None)
],
additional_inputs=[
gr.Radio(choices=['harvest', "crepe", "rmvpe", 'pm'], label=_("RVC Method", lang), value="harvest", interactive=True),
@@ -13378,49 +14090,9 @@ def create_footer():
stableaudio_interface.input_components[3].click(stop_generation, [], [], queue=False)
audioldm2_interface.input_components[4].click(stop_generation, [], [], queue=False)
- reload_button = gr.Button(_("Reload interface", lang))
close_button = gr.Button(_("Close terminal", lang))
folder_button = gr.Button(_("Outputs", lang))
- dropdowns_to_update = [
- chat_interface.input_components[4],
- chat_interface.input_components[5],
- chat_interface.input_components[6],
- chat_interface.input_components[44],
- tts_stt_interface.input_components[3],
- txt2img_interface.input_components[3],
- txt2img_interface.input_components[5],
- txt2img_interface.input_components[6],
- txt2img_interface.input_components[8],
- img2img_interface.input_components[5],
- img2img_interface.input_components[7],
- img2img_interface.input_components[8],
- img2img_interface.input_components[10],
- controlnet_interface.input_components[4],
- inpaint_interface.input_components[6],
- inpaint_interface.input_components[7],
- outpaint_interface.input_components[4],
- gligen_interface.input_components[5],
- animatediff_interface.input_components[5],
- sd3_txt2img_interface.input_components[4],
- sd3_txt2img_interface.input_components[9],
- sd3_txt2img_interface.input_components[10],
- sd3_img2img_interface.input_components[6],
- t2i_ip_adapter_interface.input_components[4],
- ip_adapter_faceid_interface.input_components[5],
- flux_txt2img_interface.input_components[2],
- flux_txt2img_interface.input_components[6],
- flux_txt2img_interface.input_components[7],
- flux_img2img_interface.input_components[3],
- flux_img2img_interface.input_components[7],
- flux_img2img_interface.input_components[8],
- auraflow_interface.input_components[3],
- kolors_txt2img_interface.input_components[3],
- rvc_interface.input_components[1],
- gallery_interface.input_components[0]
- ]
-
- reload_button.click(reload_interface, outputs=dropdowns_to_update[:16])
close_button.click(close_terminal, [], [], queue=False)
folder_button.click(open_outputs_folder, [], [], queue=False)
diff --git a/README.md b/README.md
index 214951be..1e257105 100644
--- a/README.md
+++ b/README.md
@@ -26,7 +26,7 @@ The goal of the project - to create the easiest possible application to use neur
## Features:
-* Easy installation via install.bat (Windows) or install.sh (Linux)
+* Easy installation via install.bat (Windows) or install.sh (Linux & MacOS)
* You can use the application via your mobile device in localhost (Via IPv4) or anywhere online (Via Share)
* Flexible and optimized interface (By Gradio)
* Debug logging to logs from `Install` and `Update` files
@@ -56,7 +56,7 @@ The goal of the project - to create the easiest possible application to use neur
* Model settings inside the interface
* Online and offline Wiki
* Gallery
-* ModelDownloader (For LLM and StableDiffusion)
+* ModelDownloader
* Application settings
* Ability to see system sensors
@@ -72,7 +72,7 @@ The goal of the project - to create the easiest possible application to use neur
## Minimum System Requirements:
-* System: Windows or Linux
+* System: Windows, Linux or MacOS
* GPU: 6GB+ or CPU: 8 core 3.6GHZ
* RAM: 16GB+
* Disk space: 20GB+
@@ -84,25 +84,25 @@ The goal of the project - to create the easiest possible application to use neur
1) First install all [RequiredDependencies](/#Required-Dependencies)
2) `Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git` to any location
-3) Run the `Install.bat` and wait for installation
-4) After installation, run `Start.bat`
-5) Wait for the application to launch
-6) Now you can start generating!
+3) Run the `Install.bat`, select your version and wait for installation
+4) After installation, run `Start.bat` and go through the initial setup
+5) Wait for the application to launch and follow the link from the terminal
+6) Now you can start generating. Enjoy!
-To get update, run `Update.bat`
-To work with the virtual environment through the terminal, run `Venv.bat`
+* To get update, run `Update.bat`
+* To work with the virtual environment through the terminal, run `Venv.bat`
-### Linux
+### Linux & MacOS
1) First install all [RequiredDependencies](/#Required-Dependencies)
2) `Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git` to any location
-3) In the terminal, run the `./Install.sh` and wait for installation of all dependencies
-4) After installation, run `./Start.sh`
-5) Wait for the application to launch
-6) Now you can start generating!
+3) Run the `./Install.sh`, select your version and wait for installation
+4) After installation, run `./Start.sh` and go through the initial setup
+5) Wait for the application to launch and follow the link from the terminal
+6) Now you can start generating. Enjoy!
-To get update, run `./Update.sh`
-To work with the virtual environment through the terminal, run `./Venv.sh`
+* To get update, run `./Update.sh`
+* To work with the virtual environment through the terminal, run `./Venv.sh`
## Wiki
diff --git a/RequirementsFiles/requirements-llama-cpp-CPU.txt b/RequirementsFiles/requirements-llama-cpp-CPU.txt
index 6ec647c2..9f1e9127 100644
--- a/RequirementsFiles/requirements-llama-cpp-CPU.txt
+++ b/RequirementsFiles/requirements-llama-cpp-CPU.txt
@@ -1 +1 @@
-llama-cpp-python==0.3.1
\ No newline at end of file
+llama-cpp-python==0.3.1 -C cmake.args="-DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS"
\ No newline at end of file
diff --git a/RequirementsFiles/requirements-llama-cpp-MPS.txt b/RequirementsFiles/requirements-llama-cpp-MPS.txt
new file mode 100644
index 00000000..d679483c
--- /dev/null
+++ b/RequirementsFiles/requirements-llama-cpp-MPS.txt
@@ -0,0 +1 @@
+llama-cpp-python==0.3.1 -C cmake.args="-DGGML_METAL=on"
\ No newline at end of file
diff --git a/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt b/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt
index 4c2504b2..8c2e508f 100644
--- a/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt
+++ b/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt
@@ -1 +1 @@
-stable-diffusion-cpp-python==0.2.1
\ No newline at end of file
+stable-diffusion-cpp-python==0.2.1 -C cmake.args="-DGGML_OPENBLAS=ON"
\ No newline at end of file
diff --git a/RequirementsFiles/requirements-stable-diffusion-cpp-MPS.txt b/RequirementsFiles/requirements-stable-diffusion-cpp-MPS.txt
new file mode 100644
index 00000000..e917b28f
--- /dev/null
+++ b/RequirementsFiles/requirements-stable-diffusion-cpp-MPS.txt
@@ -0,0 +1 @@
+stable-diffusion-cpp-python==0.2.1 -C cmake.args="-DSD_METAL=ON"
\ No newline at end of file
diff --git "a/RequirementsFiles/requirements-\320\241PU.txt" "b/RequirementsFiles/requirements-\320\241PU.txt"
index 39cdbfac..7b5bda6d 100644
--- "a/RequirementsFiles/requirements-\320\241PU.txt"
+++ "b/RequirementsFiles/requirements-\320\241PU.txt"
@@ -84,7 +84,7 @@ gitdb==4.0.11
GitPython==3.1.43
google-pasta==0.2.0
gpytoolbox==0.3.2
-gradio==5.4.0
+gradio==5.5.0
gradio_client==1.4.2
grpcio==1.62.2
gruut==2.2.3
@@ -407,4 +407,5 @@ stack-data==0.6.3
tornado==6.4.1
traitlets==5.14.3
widgetsnbextension==4.0.13
-rvc-python==0.1.5
\ No newline at end of file
+rvc-python==0.1.5
+tensorflow-metal==1.1.0
\ No newline at end of file
diff --git a/RequirementsFiles/requirements.txt b/RequirementsFiles/requirements.txt
index 3043f0d6..44c81727 100644
--- a/RequirementsFiles/requirements.txt
+++ b/RequirementsFiles/requirements.txt
@@ -86,7 +86,7 @@ GitPython==3.1.43
google-pasta==0.2.0
GPUtil==1.4.0
gpytoolbox==0.3.2
-gradio==5.4.0
+gradio==5.5.0
gradio_client==1.4.2
grpcio==1.62.2
gruut==2.2.3
diff --git a/TechnicalFiles/Readmes/README_RU.md b/TechnicalFiles/Readmes/README_RU.md
index 4b25a39b..3a7dd1b7 100644
--- a/TechnicalFiles/Readmes/README_RU.md
+++ b/TechnicalFiles/Readmes/README_RU.md
@@ -26,7 +26,7 @@
## Функции:
-* Легкая установка через install.bat (Windows) или install.sh (Linux)
+* Легкая установка через install.bat (Windows) или install.sh (Linux & MacOS)
* Возможность использования приложения через мобильное устройство в локальной сети (через IPv4) или в интернете (через Share)
* Гибкий и оптимизированный интерфейс (на основе Gradio)
* Ведение журнала отладки в логах из файлов `Install` и `Update`
@@ -56,7 +56,7 @@
* Настройки моделей внутри интерфейса
* Онлайн и оффлайн Wiki
* Галерея
-* ModelDownloader (Для LLM и StableDiffusion)
+* ModelDownloader
* Настройки приложения
* Возможность просмотра системных датчиков
@@ -72,7 +72,7 @@
## Минимальные системные требования:
-* Система: Windows или Linux
+* Система: Windows, Linux или MacOS
* GPU: 6GB+ или CPU: 8 ядер 3.6GHZ
* ОЗУ: 16GB+
* Место на диске: 20GB+
@@ -80,29 +80,29 @@
## Как установить:
-### Windows
+### Windows & MacOS
1) Сначала установите все [Необходимые зависимости](/#Required-Dependencies)
2) `Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git` в любое место
-3) Запустите `Install.bat` и дождитесь завершения установки
-4) После установки запустите `Start.bat`
-5) Дождитесь запуска приложения
-6) Теперь вы можете начать генерацию!
+3) Запустите `Install.bat`, выберите свою версию и дождитесь установки
+4) После установки запустите `Start.bat` и пройдите начальную настройку
+5) Дождитесь запуска приложения и перейдите по ссылке из терминала
+6) Теперь вы можете начинать генерацию. Наслаждайтесь!
-Для получения обновлений запустите `Update.bat`
-Для работы с виртуальной средой через терминал запустите `Venv.bat`
+* Для получения обновлений запустите `Update.bat`
+* Для работы с виртуальной средой через терминал запустите `Venv.bat`
### Linux
1) Сначала установите все [Необходимые зависимости](/#Required-Dependencies)
2) `Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git` в любое место
-3) В терминале запустите `./Install.sh` и дождитесь установки всех зависимостей
-4) После установки запустите `./Start.sh`
-5) Дождитесь запуска приложения
-6) Теперь вы можете начать генерацию!
+3) Запустите `./Install.sh`, выберите свою версию и дождитесь установки
+4) После установки запустите `./Start.sh` и пройдите начальную настройку
+5) Дождитесь запуска приложения и перейдите по ссылке из терминала
+6) Теперь вы можете начинать генерацию. Наслаждайтесь!
-Для получения обновлений запустите `./Update.sh`
-Для работы с виртуальной средой через терминал запустите `./Venv.sh`
+* Для получения обновлений запустите `./Update.sh`
+* Для работы с виртуальной средой через терминал запустите `./Venv.sh`
## Вики
diff --git a/TechnicalFiles/Readmes/README_ZH.md b/TechnicalFiles/Readmes/README_ZH.md
index 8cbb514d..520732ce 100644
--- a/TechnicalFiles/Readmes/README_ZH.md
+++ b/TechnicalFiles/Readmes/README_ZH.md
@@ -26,7 +26,7 @@
## 功能：
-* 通过install.bat（Windows）或install.sh（Linux）轻松安装
+* 通过install.bat（Windows）或install.sh（Linux & MacOS）轻松安装
* 您可以通过移动设备在本地主机（通过IPv4）或在线任何地方（通过Share）使用应用程序
* 灵活且优化的界面（由Gradio提供）
* 从`Install`和`Update`文件进行调试日志记录
@@ -56,7 +56,7 @@
* 界面内的模型设置
* 在线和离线Wiki
* 图库
-* 模型下载器（用于LLM和StableDiffusion）
+* 模型下载器
* 应用程序设置
* 能够查看系统传感器
@@ -72,7 +72,7 @@
## 最低系统要求：
-* 系统：Windows或Linux
+* 系统：Windows, Linux或MacOS
* GPU：6GB+或CPU：8核3.6GHZ
* RAM：16GB+
* 磁盘空间：20GB+
@@ -84,25 +84,25 @@
1) 首先安装所有[必需依赖](/#必需依赖)
2) 在任意位置执行`Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git`
-3) 运行`Install.bat`并等待安装完成
-4) 安装完成后，运行`Start.bat`
-5) 等待应用程序启动
-6) 现在您可以开始生成了！
+3) 运行 `Install.bat`, 选择您的版本并等待安装
+4) 安装完成后, 运行 `Start.bat` 并进行初始设置
+5) 等待应用程序启动并从终端中跟随链接
+6) 现在您可以开始生成了。 请尽情享用!
-要获取更新，请运行`Update.bat`
-要通过终端使用虚拟环境，请运行`Venv.bat`
+* 要获取更新，请运行`Update.bat`
+* 要通过终端使用虚拟环境，请运行`Venv.bat`
-### Linux
+### Linux & MacOS
1) 首先安装所有[必需依赖](/#必需依赖)
2) 在任意位置执行`Git clone https://github.com/Dartvauder/NeuroSandboxWebUI.git`
-3) 在终端中运行`./Install.sh`并等待所有依赖项安装完成
-4) 安装完成后，运行`./Start.sh`
-5) 等待应用程序启动
-6) 现在您可以开始生成了！
+3) 运行 `./Install.sh`, 选择您的版本并等待安装
+4) 安装完成后, 运行 `./Start.sh` 并进行初始设置
+5) 等待应用程序启动并从终端中跟随链接
+6) 现在您可以开始生成了。 请尽情享用!
-要获取更新，请运行`./Update.sh`
-要通过终端使用虚拟环境，请运行`./Venv.sh`
+* 要获取更新，请运行`./Update.sh`
+* 要通过终端使用虚拟环境，请运行`./Venv.sh`
## Wiki
diff --git a/Update.sh b/Update.sh
index d92f0930..2e687611 100644
--- a/Update.sh
+++ b/Update.sh
@@ -28,19 +28,34 @@ touch "$ERROR_LOG"
if [ "$INSTALL_TYPE" = "CPU" ]; then
export BUILD_CUDA_EXT=0
export INSTALL_KERNELS=0
+
+ if system_profiler SPDisplaysDataType | grep -q "Metal"; then
+ echo "MPS is detected. Installing MPS-specific requirements."
+ MPS_MODE=true
+ else
+ MPS_MODE=false
+ fi
else
export BUILD_CUDA_EXT=1
export INSTALL_KERNELS=1
+ MPS_MODE=false
fi
python3 -m pip install --upgrade pip setuptools
pip install wheel
if [ "$INSTALL_TYPE" = "CPU" ]; then
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-CPU.txt" 2>> "$ERROR_LOG"
- pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ if [ "$MPS_MODE" = true ]; then
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-MPS.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-MPS.txt" 2>> "$ERROR_LOG"
+ else
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-СPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-llama-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-stable-diffusion-cpp-CPU.txt" 2>> "$ERROR_LOG"
+ fi
else
pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements.txt" 2>> "$ERROR_LOG"
pip install --no-deps -r "$CURRENT_DIR/RequirementsFiles/requirements-cuda.txt" 2>> "$ERROR_LOG"