Upload 8 files

.gitattributes

@@ -1,35 +1,52 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoints/BFM_Fitting/01_MorphableModel.mat filter=lfs diff=lfs merge=lfs -text
+checkpoints/BFM_Fitting/BFM09_model_info.mat filter=lfs diff=lfs merge=lfs -text
+checkpoints/facevid2vid_00189-model.pth.tar filter=lfs diff=lfs merge=lfs -text
+checkpoints/mapping_00229-model.pth.tar filter=lfs diff=lfs merge=lfs -text
+checkpoints/shape_predictor_68_face_landmarks.dat filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/chinese_news.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/deyu.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/eluosi.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/fayu.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/imagine.wav filter=lfs diff=lfs merge=lfs -text
+examples/driven_audio/japanese.wav filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_16.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_17.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_3.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_4.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_5.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_8.png filter=lfs diff=lfs merge=lfs -text
+examples/source_image/art_9.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,159 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/#use-with-ide
+.pdm.toml
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+results/
+checkpoints/
+gradio_cached_examples/
+gfpgan/
+start.sh

Dockerfile

@@ -0,0 +1,59 @@
+FROM nvidia/cuda:11.7.1-cudnn8-devel-ubuntu22.04
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update && \
+    apt-get upgrade -y && \
+    apt-get install -y --no-install-recommends \
+    git \
+    zip \
+    unzip \
+    git-lfs \
+    wget \
+    curl \
+    # ffmpeg \
+    ffmpeg \
+    x264 \
+    # python build dependencies \
+    build-essential \
+    libssl-dev \
+    zlib1g-dev \
+    libbz2-dev \
+    libreadline-dev \
+    libsqlite3-dev \
+    libncursesw5-dev \
+    xz-utils \
+    tk-dev \
+    libxml2-dev \
+    libxmlsec1-dev \
+    libffi-dev \
+    liblzma-dev && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:${PATH}
+WORKDIR ${HOME}/app
+
+RUN curl https://pyenv.run | bash
+ENV PATH=${HOME}/.pyenv/shims:${HOME}/.pyenv/bin:${PATH}
+ENV PYTHON_VERSION=3.10.9
+RUN pyenv install ${PYTHON_VERSION} && \
+    pyenv global ${PYTHON_VERSION} && \
+    pyenv rehash && \
+    pip install --no-cache-dir -U pip setuptools wheel
+
+RUN pip install --no-cache-dir -U torch==1.12.1 torchvision==0.13.1
+COPY --chown=1000 requirements.txt /tmp/requirements.txt
+RUN pip install --no-cache-dir -U -r /tmp/requirements.txt
+
+COPY --chown=1000 . ${HOME}/app
+RUN ls -a
+ENV PYTHONPATH=${HOME}/app \
+    PYTHONUNBUFFERED=1 \
+    GRADIO_ALLOW_FLAGGING=never \
+    GRADIO_NUM_PORTS=1 \
+    GRADIO_SERVER_NAME=0.0.0.0 \
+    GRADIO_THEME=huggingface \
+    SYSTEM=spaces
+CMD ["python", "app.py"]

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2023 Tencent AI Lab
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,13 +1,12 @@
----
-title: Sadtalker
-emoji: 🌖
-colorFrom: yellow
-colorTo: indigo
-sdk: gradio
-sdk_version: 4.44.1
-app_file: app.py
-pinned: false
-license: apache-2.0
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: SadTalker
+emoji: 🏢
+colorFrom: indigo
+colorTo: gray
+sdk: gradio
+sdk_version: 4.31.0
+app_file: app.py
+pinned: false
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,217 @@
+import os
+import platform
+import uuid
+import shutil
+from pydub import AudioSegment
+import spaces
+import torch
+import gradio as gr
+from huggingface_hub import snapshot_download
+
+from examples.get_examples import get_examples
+from src.facerender.pirender_animate import AnimateFromCoeff_PIRender
+from src.utils.preprocess import CropAndExtract
+from src.test_audio2coeff import Audio2Coeff
+from src.facerender.animate import AnimateFromCoeff
+from src.generate_batch import get_data
+from src.generate_facerender_batch import get_facerender_data
+from src.utils.init_path import init_path
+
+checkpoint_path = 'checkpoints'
+config_path = 'src/config'
+device = "cuda" if torch.cuda.is_available(
+) else "mps" if platform.system() == 'Darwin' else "cpu"
+
+os.environ['TORCH_HOME'] = checkpoint_path
+snapshot_download(repo_id='vinthony/SadTalker-V002rc',
+                  local_dir=checkpoint_path, local_dir_use_symlinks=True)
+
+
+def mp3_to_wav(mp3_filename, wav_filename, frame_rate):
+    AudioSegment.from_file(file=mp3_filename).set_frame_rate(
+        frame_rate).export(wav_filename, format="wav")
+
+
+@spaces.GPU(duration=120)
+def generate_video(source_image, driven_audio, preprocess='crop', still_mode=False, use_enhancer=False,
+                   batch_size=1, size=256, pose_style=0, facerender='facevid2vid', exp_scale=1.0,
+                   use_ref_video=False, ref_video=None, ref_info=None, use_idle_mode=False,
+                   length_of_audio=0, use_blink=True, result_dir='./results/'):
+    # Initialize models and paths
+    sadtalker_paths = init_path(
+        checkpoint_path, config_path, size, False, preprocess)
+    audio_to_coeff = Audio2Coeff(sadtalker_paths, device)
+    preprocess_model = CropAndExtract(sadtalker_paths, device)
+    animate_from_coeff = AnimateFromCoeff(sadtalker_paths, device) if facerender == 'facevid2vid' and device != 'mps' \
+        else AnimateFromCoeff_PIRender(sadtalker_paths, device)
+
+    # Create directories for saving results
+    time_tag = str(uuid.uuid4())
+    save_dir = os.path.join(result_dir, time_tag)
+    os.makedirs(save_dir, exist_ok=True)
+    input_dir = os.path.join(save_dir, 'input')
+    os.makedirs(input_dir, exist_ok=True)
+
+    # Process source image
+    pic_path = os.path.join(input_dir, os.path.basename(source_image))
+    shutil.move(source_image, input_dir)
+
+    # Process driven audio
+    if driven_audio and os.path.isfile(driven_audio):
+        audio_path = os.path.join(input_dir, os.path.basename(driven_audio))
+        if '.mp3' in audio_path:
+            mp3_to_wav(driven_audio, audio_path.replace('.mp3', '.wav'), 16000)
+            audio_path = audio_path.replace('.mp3', '.wav')
+        else:
+            shutil.move(driven_audio, input_dir)
+    elif use_idle_mode:
+        audio_path = os.path.join(
+            input_dir, 'idlemode_'+str(length_of_audio)+'.wav')
+        AudioSegment.silent(
+            duration=1000*length_of_audio).export(audio_path, format="wav")
+    else:
+        assert use_ref_video and ref_info == 'all'
+
+    # Process reference video
+    if use_ref_video and ref_info == 'all':
+        ref_video_videoname = os.path.splitext(os.path.split(ref_video)[-1])[0]
+        audio_path = os.path.join(save_dir, ref_video_videoname+'.wav')
+        os.system(
+            f"ffmpeg -y -hide_banner -loglevel error -i {ref_video} {audio_path}")
+        ref_video_frame_dir = os.path.join(save_dir, ref_video_videoname)
+        os.makedirs(ref_video_frame_dir, exist_ok=True)
+        ref_video_coeff_path, _, _ = preprocess_model.generate(
+            ref_video, ref_video_frame_dir, preprocess, source_image_flag=False)
+    else:
+        ref_video_coeff_path = None
+
+    # Preprocess source image
+    first_frame_dir = os.path.join(save_dir, 'first_frame_dir')
+    os.makedirs(first_frame_dir, exist_ok=True)
+    first_coeff_path, crop_pic_path, crop_info = preprocess_model.generate(
+        pic_path, first_frame_dir, preprocess, True, size)
+    if first_coeff_path is None:
+        raise AttributeError("No face is detected")
+
+    # Determine reference coefficients
+    ref_pose_coeff_path, ref_eyeblink_coeff_path = None, None
+    if use_ref_video:
+        if ref_info == 'pose':
+            ref_pose_coeff_path = ref_video_coeff_path
+        elif ref_info == 'blink':
+            ref_eyeblink_coeff_path = ref_video_coeff_path
+        elif ref_info == 'pose+blink':
+            ref_pose_coeff_path = ref_eyeblink_coeff_path = ref_video_coeff_path
+    else:
+        ref_pose_coeff_path = ref_eyeblink_coeff_path = None
+
+    # Generate coefficients from audio or reference video
+    if use_ref_video and ref_info == 'all':
+        coeff_path = ref_video_coeff_path
+    else:
+        batch = get_data(first_coeff_path, audio_path, device, ref_eyeblink_coeff_path=ref_eyeblink_coeff_path,
+                         still=still_mode, idlemode=use_idle_mode, length_of_audio=length_of_audio, use_blink=use_blink)
+        coeff_path = audio_to_coeff.generate(
+            batch, save_dir, pose_style, ref_pose_coeff_path)
+
+    # Generate video from coefficients
+    data = get_facerender_data(coeff_path, crop_pic_path, first_coeff_path, audio_path, batch_size, still_mode=still_mode,
+                               preprocess=preprocess, size=size, expression_scale=exp_scale, facemodel=facerender)
+    return_path = animate_from_coeff.generate(data, save_dir, pic_path, crop_info, enhancer='gfpgan' if use_enhancer else None,
+                                              preprocess=preprocess, img_size=size)
+    video_name = data['video_name']
+    print(f'The generated video is named {video_name} in {save_dir}')
+
+    return return_path
+
+
+# Gradio UI
+with gr.Blocks(analytics_enabled=False) as demo:
+    with gr.Row():
+        with gr.Column(variant='panel'):
+            with gr.Tabs(elem_id="sadtalker_source_image"):
+                with gr.TabItem('Source image'):
+                    with gr.Row():
+                        source_image = gr.Image(
+                            label="Source image", sources="upload", type="filepath", elem_id="img2img_image")
+
+            with gr.Tabs(elem_id="sadtalker_driven_audio"):
+                with gr.TabItem('Driving Methods'):
+                    gr.Markdown(
+                        "Possible driving combinations: <br> 1. Audio only 2. Audio/IDLE Mode + Ref Video(pose, blink, pose+blink) 3. IDLE Mode only 4. Ref Video only (all) ")
+
+                    with gr.Row():
+                        driven_audio = gr.Audio(
+                            label="Input audio", sources="upload", type="filepath")
+                        driven_audio_no = gr.Audio(
+                            label="Use IDLE mode, no audio is required", sources="upload", type="filepath", visible=False)
+
+                        with gr.Column():
+                            use_idle_mode = gr.Checkbox(
+                                label="Use Idle Animation")
+                            length_of_audio = gr.Number(
+                                value=5, label="The length(seconds) of the generated video.")
+                            use_idle_mode.change(lambda choice: (gr.update(visible=not choice), gr.update(visible=choice)),
+                                                 inputs=use_idle_mode, outputs=[driven_audio, driven_audio_no])
+
+                    with gr.Row():
+                        ref_video = gr.Video(
+                            label="Reference Video", sources="upload", elem_id="vidref")
+
+                        with gr.Column():
+                            use_ref_video = gr.Checkbox(
+                                label="Use Reference Video")
+                            ref_info = gr.Radio(['pose', 'blink', 'pose+blink', 'all'], value='pose', label='Reference Video',
+                                                info="How to borrow from reference Video?((fully transfer, aka, video driving mode))")
+
+                        ref_video.change(lambda path: gr.update(
+                            value=path is not None), inputs=ref_video, outputs=use_ref_video)
+
+        with gr.Column(variant='panel'):
+            with gr.Tabs(elem_id="sadtalker_checkbox"):
+                with gr.TabItem('Settings'):
+                    with gr.Column(variant='panel'):
+                        with gr.Row():
+                            pose_style = gr.Slider(
+                                minimum=0, maximum=45, step=1, label="Pose style", value=0)
+                            exp_weight = gr.Slider(
+                                minimum=0, maximum=3, step=0.1, label="expression scale", value=1)
+                            blink_every = gr.Checkbox(
+                                label="use eye blink", value=True)
+
+                        with gr.Row():
+                            size_of_image = gr.Radio(
+                                [256, 512], value=256, label='face model resolution', info="use 256/512 model?")
+                            preprocess_type = gr.Radio(
+                                ['crop', 'resize', 'full', 'extcrop', 'extfull'], value='crop', label='preprocess', info="How to handle input image?")
+
+                        with gr.Row():
+                            is_still_mode = gr.Checkbox(
+                                label="Still Mode (fewer head motion, works with preprocess `full`)")
+                            facerender = gr.Radio(
+                                ['facevid2vid', 'pirender'], value='facevid2vid', label='facerender', info="which face render?")
+
+                        with gr.Row():
+                            batch_size = gr.Slider(
+                                label="batch size in generation", step=1, maximum=10, value=1)
+                            enhancer = gr.Checkbox(
+                                label="GFPGAN as Face enhancer", value=True)
+
+                        submit = gr.Button(
+                            'Generate', elem_id="sadtalker_generate", variant='primary')
+
+            with gr.Tabs(elem_id="sadtalker_generated"):
+                gen_video = gr.Video(label="Generated video")
+
+    submit.click(
+        fn=generate_video,
+        inputs=[source_image, driven_audio, preprocess_type, is_still_mode, enhancer, batch_size, size_of_image,
+                pose_style, facerender, exp_weight, use_ref_video, ref_video, ref_info, use_idle_mode, length_of_audio, blink_every],
+        outputs=[gen_video],
+    )
+
+    with gr.Row():
+        gr.Examples(examples=get_examples(), inputs=[source_image, driven_audio, preprocess_type, is_still_mode, enhancer],
+                    outputs=[gen_video], fn=generate_video)
+
+demo.launch(debug=True)

packages.txt

@@ -0,0 +1,2 @@
+ffmpeg
+libsndfile1

requirements.txt

@@ -0,0 +1,24 @@
+torch
+torchvision==0.15.2
+torchaudio
+numpy==1.23.5
+face_alignment
+imageio
+imageio-ffmpeg
+librosa
+numba
+resampy
+pydub
+scipy
+kornia
+tqdm
+yacs
+pyyaml  
+joblib
+scikit-image
+basicsr
+facexlib
+dlib-bin
+gfpgan
+av
+safetensors