Upload folder using huggingface_hub

.gitattributes

@@ -1,35 +1,38 @@
-*.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
+*.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
+*.png filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text
+shape_predictor_68_face_landmarks.dat filter=lfs diff=lfs merge=lfs -text

=2.0.0

@@ -0,0 +1,10 @@
+Requirement already satisfied: grpcio in c:\users\administrator\appdata\local\programs\python\python310\lib\site-packages (1.63.0)
+Collecting grpcio
+  Using cached grpcio-1.64.1-cp310-cp310-win_amd64.whl.metadata (3.4 kB)
+Using cached grpcio-1.64.1-cp310-cp310-win_amd64.whl (4.1 MB)
+Installing collected packages: grpcio
+  Attempting uninstall: grpcio
+    Found existing installation: grpcio 1.63.0
+    Uninstalling grpcio-1.63.0:
+      Successfully uninstalled grpcio-1.63.0
+Successfully installed grpcio-1.64.1

README.md

@@ -1,12 +1,17 @@
----
-title: HairFastGAN
-emoji: 🐨
-colorFrom: blue
-colorTo: indigo
-sdk: gradio
-sdk_version: 4.36.1
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: HairFastGAN
+emoji: 💈
+colorFrom: pink
+colorTo: blue
+sdk: gradio
+sdk_version: 4.31.5
+app_file: app.py
+pinned: false
+license: mit
+custom_headers:
+  cross-origin-embedder-policy: require-corp
+  cross-origin-opener-policy: same-origin
+  cross-origin-resource-policy: cross-origin
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,156 @@
+import hashlib
+import os
+from io import BytesIO
+
+import gradio as gr
+import grpc
+from PIL import Image
+from cachetools import LRUCache
+
+from inference_pb2 import HairSwapRequest, HairSwapResponse
+from inference_pb2_grpc import HairSwapServiceStub
+from utils.shape_predictor import align_face
+
+
+def get_bytes(img):
+    if img is None:
+        return img
+
+    buffered = BytesIO()
+    img.save(buffered, format="JPEG")
+    return buffered.getvalue()
+
+
+def bytes_to_image(image: bytes) -> Image.Image:
+    image = Image.open(BytesIO(image))
+    return image
+
+
+def center_crop(img):
+    width, height = img.size
+    side = min(width, height)
+
+    left = (width - side) / 2
+    top = (height - side) / 2
+    right = (width + side) / 2
+    bottom = (height + side) / 2
+
+    img = img.crop((left, top, right, bottom))
+    return img
+
+
+def resize(name):
+    def resize_inner(img, align):
+        global align_cache
+
+        if name in align:
+            img_hash = hashlib.md5(get_bytes(img)).hexdigest()
+
+            if img_hash not in align_cache:
+                img = align_face(img, return_tensors=False)[0]
+                align_cache[img_hash] = img
+            else:
+                img = align_cache[img_hash]
+
+        elif img.size != (1024, 1024):
+            img = center_crop(img)
+            img = img.resize((1024, 1024), Image.Resampling.LANCZOS)
+
+        return img
+
+    return resize_inner
+
+
+def swap_hair(face, shape, color, blending, poisson_iters, poisson_erosion):
+    if not face and not shape and not color:
+        return gr.update(visible=False), gr.update(value="Need to upload a face and at least a shape or color ❗", visible=True)
+    elif not face:
+        return gr.update(visible=False), gr.update(value="Need to upload a face ❗", visible=True)
+    elif not shape and not color:
+        return gr.update(visible=False), gr.update(value="Need to upload at least a shape or color ❗", visible=True)
+
+    face_bytes, shape_bytes, color_bytes = map(lambda item: get_bytes(item), (face, shape, color))
+
+    if shape_bytes is None:
+        shape_bytes = b'face'
+    if color_bytes is None:
+        color_bytes = b'shape'
+    if os.environ.get('https_proxy'):
+        del os.environ['https_proxy']
+    if os.environ.get('http_proxy'):
+        del os.environ['http_proxy']
+    os.environ['SERVER'] = '172.16.4.26:7860'
+    
+    # with grpc.insecure_channel(os.environ['SERVER'], options=(('grpc.enable_http_proxy', 0),)) as channel:
+    #     stub = HairSwapServiceStub(channel)
+
+    #     output: HairSwapResponse = stub.swap(
+    #         HairSwapRequest(face=face_bytes, shape=shape_bytes, color=color_bytes, blending=blending,
+    #                         poisson_iters=poisson_iters, poisson_erosion=poisson_erosion, use_cache=True)
+    #     )
+
+    # output = bytes_to_image(output.image)
+    # return gr.update(value=output, visible=True), gr.update(visible=False)
+
+
+def get_demo():
+    with gr.Blocks() as demo:
+        gr.Markdown("## HairFastGan")
+        gr.Markdown(
+            '<div style="display: flex; align-items: center; gap: 10px;">'
+            '<span>Official HairFastGAN Gradio demo:</span>'
+            '<a href="https://arxiv.org/abs/2404.01094"><img src="https://img.shields.io/badge/arXiv-2404.01094-b31b1b.svg" height=22.5></a>'
+            '<a href="https://github.com/AIRI-Institute/HairFastGAN"><img src="https://img.shields.io/badge/github-%23121011.svg?style=for-the-badge&logo=github&logoColor=white" height=22.5></a>'
+            '<a href="https://huggingface.co/AIRI-Institute/HairFastGAN"><img src="https://huggingface.co/datasets/huggingface/badges/resolve/main/model-on-hf-md.svg" height=22.5></a>'
+            '<a href="https://colab.research.google.com/#fileId=https://huggingface.co/AIRI-Institute/HairFastGAN/blob/main/notebooks/HairFast_inference.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" height=22.5></a>'
+            '</div>'
+        )
+        with gr.Row():
+            with gr.Column():
+                source = gr.Image(label="Source photo to try on the hairstyle", type="pil")
+                with gr.Row():
+                    shape = gr.Image(label="Shape photo with desired hairstyle (optional)", type="pil")
+                    color = gr.Image(label="Color photo with desired hair color (optional)", type="pil")
+                with gr.Accordion("Advanced Options", open=False):
+                    blending = gr.Radio(["Article", "Alternative_v1", "Alternative_v2"], value='Article',
+                                        label="Color Encoder version", info="Selects a model for hair color transfer.")
+                    poisson_iters = gr.Slider(0, 2500, value=0, step=1, label="Poisson iters",
+                                              info="The power of blending with the original image, helps to recover more details. Not included in the article, disabled by default.")
+                    poisson_erosion = gr.Slider(1, 100, value=15, step=1, label="Poisson erosion",
+                                                info="Smooths out the blending area.")
+                    align = gr.CheckboxGroup(["Face", "Shape", "Color"], value=["Face", "Shape", "Color"],
+                                             label="Image cropping [recommended]",
+                                             info="Selects which images to crop by face")
+                btn = gr.Button("Get the haircut")
+            with gr.Column():
+                output = gr.Image(label="Your result")
+                error_message = gr.Textbox(label="⚠️ Error ⚠️", visible=False, elem_classes="error-message")
+
+        gr.Examples(examples=[["input/0.png", "input/1.png", "input/2.png"], ["input/6.png", "input/7.png", None],
+                              ["input/10.jpg", None, "input/11.jpg"]],
+                    inputs=[source, shape, color], outputs=output)
+
+        source.upload(fn=resize('Face'), inputs=[source, align], outputs=source)
+        shape.upload(fn=resize('Shape'), inputs=[shape, align], outputs=shape)
+        color.upload(fn=resize('Color'), inputs=[color, align], outputs=color)
+
+        btn.click(fn=swap_hair, inputs=[source, shape, color, blending, poisson_iters, poisson_erosion],
+                  outputs=[output, error_message])
+
+        gr.Markdown('''To cite the paper by the authors
+    ```
+        @article{nikolaev2024hairfastgan,
+          title={HairFastGAN: Realistic and Robust Hair Transfer with a Fast Encoder-Based Approach},
+          author={Nikolaev, Maxim and Kuznetsov, Mikhail and Vetrov, Dmitry and Alanov, Aibek},
+          journal={arXiv preprint arXiv:2404.01094},
+          year={2024}
+        }
+    ```
+        ''')
+    return demo
+
+
+if __name__ == '__main__':
+    align_cache = LRUCache(maxsize=10)
+    demo = get_demo()
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=True)

inference_pb2.py

@@ -0,0 +1,30 @@
+# -*- coding: utf-8 -*-
+# Generated by the protocol buffer compiler.  DO NOT EDIT!
+# source: inference.proto
+# Protobuf Python Version: 5.26.1
+"""Generated protocol buffer code."""
+from google.protobuf import descriptor as _descriptor
+from google.protobuf import descriptor_pool as _descriptor_pool
+from google.protobuf import symbol_database as _symbol_database
+from google.protobuf.internal import builder as _builder
+# @@protoc_insertion_point(imports)
+
+_sym_db = _symbol_database.Default()
+
+
+
+
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x0finference.proto\x12\tinference\"\x92\x01\n\x0fHairSwapRequest\x12\x0c\n\x04\x66\x61\x63\x65\x18\x01 \x01(\x0c\x12\r\n\x05shape\x18\x02 \x01(\x0c\x12\r\n\x05\x63olor\x18\x03 \x01(\x0c\x12\x10\n\x08\x62lending\x18\x04 \x01(\t\x12\x15\n\rpoisson_iters\x18\x05 \x01(\x05\x12\x17\n\x0fpoisson_erosion\x18\x06 \x01(\x05\x12\x11\n\tuse_cache\x18\x07 \x01(\x08\"!\n\x10HairSwapResponse\x12\r\n\x05image\x18\x01 \x01(\x0c\x32R\n\x0fHairSwapService\x12?\n\x04swap\x12\x1a.inference.HairSwapRequest\x1a\x1b.inference.HairSwapResponseb\x06proto3')
+
+_globals = globals()
+_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
+_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'inference_pb2', _globals)
+if not _descriptor._USE_C_DESCRIPTORS:
+  DESCRIPTOR._loaded_options = None
+  _globals['_HAIRSWAPREQUEST']._serialized_start=31
+  _globals['_HAIRSWAPREQUEST']._serialized_end=177
+  _globals['_HAIRSWAPRESPONSE']._serialized_start=179
+  _globals['_HAIRSWAPRESPONSE']._serialized_end=212
+  _globals['_HAIRSWAPSERVICE']._serialized_start=214
+  _globals['_HAIRSWAPSERVICE']._serialized_end=296
+# @@protoc_insertion_point(module_scope)

inference_pb2.pyi

@@ -0,0 +1,29 @@
+from google.protobuf import descriptor as _descriptor
+from google.protobuf import message as _message
+from typing import ClassVar as _ClassVar, Optional as _Optional
+
+DESCRIPTOR: _descriptor.FileDescriptor
+
+class HairSwapRequest(_message.Message):
+    __slots__ = ("face", "shape", "color", "blending", "poisson_iters", "poisson_erosion", "use_cache")
+    FACE_FIELD_NUMBER: _ClassVar[int]
+    SHAPE_FIELD_NUMBER: _ClassVar[int]
+    COLOR_FIELD_NUMBER: _ClassVar[int]
+    BLENDING_FIELD_NUMBER: _ClassVar[int]
+    POISSON_ITERS_FIELD_NUMBER: _ClassVar[int]
+    POISSON_EROSION_FIELD_NUMBER: _ClassVar[int]
+    USE_CACHE_FIELD_NUMBER: _ClassVar[int]
+    face: bytes
+    shape: bytes
+    color: bytes
+    blending: str
+    poisson_iters: int
+    poisson_erosion: int
+    use_cache: bool
+    def __init__(self, face: _Optional[bytes] = ..., shape: _Optional[bytes] = ..., color: _Optional[bytes] = ..., blending: _Optional[str] = ..., poisson_iters: _Optional[int] = ..., poisson_erosion: _Optional[int] = ..., use_cache: bool = ...) -> None: ...
+
+class HairSwapResponse(_message.Message):
+    __slots__ = ("image",)
+    IMAGE_FIELD_NUMBER: _ClassVar[int]
+    image: bytes
+    def __init__(self, image: _Optional[bytes] = ...) -> None: ...

inference_pb2_grpc.py

@@ -0,0 +1,101 @@
+# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
+"""Client and server classes corresponding to protobuf-defined services."""
+import grpc
+import warnings
+
+import inference_pb2 as inference__pb2
+
+GRPC_GENERATED_VERSION = '1.63.0'
+GRPC_VERSION = grpc.__version__
+EXPECTED_ERROR_RELEASE = '1.65.0'
+SCHEDULED_RELEASE_DATE = 'June 25, 2024'
+_version_not_supported = False
+
+try:
+    from grpc._utilities import first_version_is_lower
+    _version_not_supported = first_version_is_lower(GRPC_VERSION, GRPC_GENERATED_VERSION)
+except ImportError:
+    _version_not_supported = True
+
+if _version_not_supported:
+    warnings.warn(
+        f'The grpc package installed is at version {GRPC_VERSION},'
+        + f' but the generated code in inference_pb2_grpc.py depends on'
+        + f' grpcio>={GRPC_GENERATED_VERSION}.'
+        + f' Please upgrade your grpc module to grpcio>={GRPC_GENERATED_VERSION}'
+        + f' or downgrade your generated code using grpcio-tools<={GRPC_VERSION}.'
+        + f' This warning will become an error in {EXPECTED_ERROR_RELEASE},'
+        + f' scheduled for release on {SCHEDULED_RELEASE_DATE}.',
+        RuntimeWarning
+    )
+
+
+class HairSwapServiceStub(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def __init__(self, channel):
+        """Constructor.
+
+        Args:
+            channel: A grpc.Channel.
+        """
+        self.swap = channel.unary_unary(
+                '/inference.HairSwapService/swap',
+                request_serializer=inference__pb2.HairSwapRequest.SerializeToString,
+                response_deserializer=inference__pb2.HairSwapResponse.FromString,
+                _registered_method=True)
+
+
+class HairSwapServiceServicer(object):
+    """Missing associated documentation comment in .proto file."""
+
+    def swap(self, request, context):
+        """Missing associated documentation comment in .proto file."""
+        context.set_code(grpc.StatusCode.UNIMPLEMENTED)
+        context.set_details('Method not implemented!')
+        raise NotImplementedError('Method not implemented!')
+
+
+def add_HairSwapServiceServicer_to_server(servicer, server):
+    rpc_method_handlers = {
+            'swap': grpc.unary_unary_rpc_method_handler(
+                    servicer.swap,
+                    request_deserializer=inference__pb2.HairSwapRequest.FromString,
+                    response_serializer=inference__pb2.HairSwapResponse.SerializeToString,
+            ),
+    }
+    generic_handler = grpc.method_handlers_generic_handler(
+            'inference.HairSwapService', rpc_method_handlers)
+    server.add_generic_rpc_handlers((generic_handler,))
+
+
+ # This class is part of an EXPERIMENTAL API.
+class HairSwapService(object):
+    """Missing associated documentation comment in .proto file."""
+
+    @staticmethod
+    def swap(request,
+            target,
+            options=(),
+            channel_credentials=None,
+            call_credentials=None,
+            insecure=False,
+            compression=None,
+            wait_for_ready=None,
+            timeout=None,
+            metadata=None):
+        return grpc.experimental.unary_unary(
+            request,
+            target,
+            '/inference.HairSwapService/swap',
+            inference__pb2.HairSwapRequest.SerializeToString,
+            inference__pb2.HairSwapResponse.FromString,
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+            _registered_method=True)

requirements.txt

@@ -0,0 +1,10 @@
+pillow==10.0.0
+face_alignment==1.3.4
+addict==2.4.0
+git+https://github.com/openai/CLIP.git
+gdown==3.12.2
+grpcio==1.63.0
+grpcio_tools==1.63.0
+gradio==4.31.5
+cachetools==5.3.3
+dlib==19.24.1

shape_predictor_68_face_landmarks.dat

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbdc2cb80eb9aa7a758672cbfdda32ba6300efe9b6e6c7a299ff7e736b11b92f
+size 99693937

test.py

@@ -0,0 +1,5 @@
+import os
+
+# Provide a default value if SERVER is not set
+server_address = os.environ.get('SERVER', '127.0.0.1')
+print(f"Server address: {os.environ['SERVER']}")

utils/drive.py

@@ -0,0 +1,110 @@
+# URL helpers, see https://github.com/NVlabs/stylegan
+# ------------------------------------------------------------------------------------------
+
+import requests
+import html
+import hashlib
+import gdown
+import glob
+import os
+import io
+from typing import Any
+import re
+import uuid
+
+weight_dic = {'afhqwild.pt': 'https://drive.google.com/file/d/14OnzO4QWaAytKXVqcfWo_o2MzoR4ygnr/view?usp=sharing',
+              'afhqdog.pt': 'https://drive.google.com/file/d/16v6jPtKVlvq8rg2Sdi3-R9qZEVDgvvEA/view?usp=sharing',
+              'afhqcat.pt': 'https://drive.google.com/file/d/1HXLER5R3EMI8DSYDBZafoqpX4EtyOf2R/view?usp=sharing',
+              'ffhq.pt': 'https://drive.google.com/file/d/1AT6bNR2ppK8f2ETL_evT27f3R_oyWNHS/view?usp=sharing',
+              'metfaces.pt': 'https://drive.google.com/file/d/16wM2PwVWzaMsRgPExvRGsq6BWw_muKbf/view?usp=sharing',
+              'seg.pth': 'https://drive.google.com/file/d/1lIKvQaFKHT5zC7uS4p17O9ZpfwmwlS62/view?usp=sharing'}
+
+
+def download_weight(weight_path):
+    gdown.download(weight_dic[os.path.basename(weight_path)],
+                   output=weight_path, fuzzy=True)
+
+
+def is_url(obj: Any) -> bool:
+    """Determine whether the given object is a valid URL string."""
+    if not isinstance(obj, str) or not "://" in obj:
+        return False
+    try:
+        res = requests.compat.urlparse(obj)
+        if not res.scheme or not res.netloc or not "." in res.netloc:
+            return False
+        res = requests.compat.urlparse(requests.compat.urljoin(obj, "/"))
+        if not res.scheme or not res.netloc or not "." in res.netloc:
+            return False
+    except:
+        return False
+    return True
+
+
+def open_url(url: str, cache_dir: str = None, num_attempts: int = 10, verbose: bool = True,
+             return_path: bool = False) -> Any:
+    """Download the given URL and return a binary-mode file object to access the data."""
+    assert is_url(url)
+    assert num_attempts >= 1
+
+    # Lookup from cache.
+    url_md5 = hashlib.md5(url.encode("utf-8")).hexdigest()
+    if cache_dir is not None:
+        cache_files = glob.glob(os.path.join(cache_dir, url_md5 + "_*"))
+        if len(cache_files) == 1:
+            if (return_path):
+                return cache_files[0]
+            else:
+                return open(cache_files[0], "rb")
+
+    # Download.
+    url_name = None
+    url_data = None
+    with requests.Session() as session:
+        if verbose:
+            print("Downloading %s ..." % url, end="", flush=True)
+        for attempts_left in reversed(range(num_attempts)):
+            try:
+                with session.get(url) as res:
+                    res.raise_for_status()
+                    if len(res.content) == 0:
+                        raise IOError("No data received")
+
+                    if len(res.content) < 8192:
+                        content_str = res.content.decode("utf-8")
+                        if "download_warning" in res.headers.get("Set-Cookie", ""):
+                            links = [html.unescape(link) for link in content_str.split('"') if
+                                     "export=download" in link]
+                            if len(links) == 1:
+                                url = requests.compat.urljoin(url, links[0])
+                                raise IOError("Google Drive virus checker nag")
+                        if "Google Drive - Quota exceeded" in content_str:
+                            raise IOError("Google Drive quota exceeded")
+
+                    match = re.search(r'filename="([^"]*)"', res.headers.get("Content-Disposition", ""))
+                    url_name = match[1] if match else url
+                    url_data = res.content
+                    if verbose:
+                        print(" done")
+                    break
+            except:
+                if not attempts_left:
+                    if verbose:
+                        print(" failed")
+                    raise
+                if verbose:
+                    print(".", end="", flush=True)
+
+    # Save to cache.
+    if cache_dir is not None:
+        safe_name = re.sub(r"[^0-9a-zA-Z-._]", "_", url_name)
+        cache_file = os.path.join(cache_dir, url_md5 + "_" + safe_name)
+        temp_file = os.path.join(cache_dir, "tmp_" + uuid.uuid4().hex + "_" + url_md5 + "_" + safe_name)
+        os.makedirs(cache_dir, exist_ok=True)
+        with open(temp_file, "wb") as f:
+            f.write(url_data)
+        os.replace(temp_file, cache_file)  # atomic
+        if (return_path): return cache_file
+
+    # Return data as file object.
+    return io.BytesIO(url_data)

utils/shape_predictor.py

@@ -0,0 +1,194 @@
+import os
+from pathlib import Path
+
+import PIL
+import dlib
+import numpy as np
+import scipy
+import scipy.ndimage
+import torch
+from PIL import Image
+from torchvision import transforms as T
+
+from utils.drive import open_url
+
+"""
+brief: face alignment with FFHQ method (https://github.com/NVlabs/ffhq-dataset)
+author: lzhbrian (https://lzhbrian.me)
+date: 2020.1.5
+note: code is heavily borrowed from
+    https://github.com/NVlabs/ffhq-dataset
+    http://dlib.net/face_landmark_detection.py.html
+
+requirements:
+    apt install cmake
+    conda install Pillow numpy scipy
+    pip install dlib
+    # download face landmark model from:
+    # http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2
+"""
+
+
+def get_landmark(filepath, predictor):
+    """get landmark with dlib
+    :return: np.array shape=(68, 2)
+    """
+    detector = dlib.get_frontal_face_detector()
+
+    img = dlib.load_rgb_image(filepath)
+    dets = detector(img, 1)
+    filepath = Path(filepath)
+    print(f"{filepath.name}: Number of faces detected: {len(dets)}")
+    shapes = [predictor(img, d) for k, d in enumerate(dets)]
+
+    lms = [np.array([[tt.x, tt.y] for tt in shape.parts()]) for shape in shapes]
+
+    return lms
+
+
+def get_landmark_from_tensors(tensors: list[torch.Tensor | Image.Image | np.ndarray], predictor):
+    detector = dlib.get_frontal_face_detector()
+    transform = T.ToPILImage()
+    images = []
+    lms = []
+
+    for k, tensor in enumerate(tensors):
+        if isinstance(tensor, torch.Tensor):
+            img_pil = transform(tensor)
+        else:
+            img_pil = tensor
+        img = np.array(img_pil)
+        images.append(img_pil)
+
+        dets = detector(img, 1)
+        if len(dets) == 0:
+            raise ValueError(f"No faces detected in the image {k}.")
+        elif len(dets) == 1:
+            print(f"Number of faces detected: {len(dets)}")
+        else:
+            print(f"Number of faces detected: {len(dets)}, get largest face")
+
+        # Find the largest face
+        dets = sorted(dets, key=lambda det: det.width() * det.height(), reverse=True)
+        shape = predictor(img, dets[0])
+        lm = np.array([[tt.x, tt.y] for tt in shape.parts()])
+        lms.append(lm)
+
+    return images, lms
+
+
+def align_face(data, predictor=None, is_filepath=False, return_tensors=True):
+    """
+    :param data: filepath or list torch Tensors
+    :return: list of PIL Images
+    """
+    if predictor is None:
+        predictor_path = 'shape_predictor_68_face_landmarks.dat'
+
+        if not os.path.isfile(predictor_path):
+            print("Downloading Shape Predictor")
+            data_io = open_url("https://drive.google.com/uc?id=1huhv8PYpNNKbGCLOaYUjOgR1pY5pmbJx")
+            with open(predictor_path, 'wb') as f:
+                f.write(data_io.getbuffer())
+
+        predictor = dlib.shape_predictor(predictor_path)
+
+    if is_filepath:
+        lms = get_landmark(data, predictor)
+    else:
+        if not isinstance(data, list):
+            data = [data]
+        images, lms = get_landmark_from_tensors(data, predictor)
+
+    imgs = []
+    for num_img, lm in enumerate(lms):
+        lm_chin = lm[0: 17]  # left-right
+        lm_eyebrow_left = lm[17: 22]  # left-right
+        lm_eyebrow_right = lm[22: 27]  # left-right
+        lm_nose = lm[27: 31]  # top-down
+        lm_nostrils = lm[31: 36]  # top-down
+        lm_eye_left = lm[36: 42]  # left-clockwise
+        lm_eye_right = lm[42: 48]  # left-clockwise
+        lm_mouth_outer = lm[48: 60]  # left-clockwise
+        lm_mouth_inner = lm[60: 68]  # left-clockwise
+
+        # Calculate auxiliary vectors.
+        eye_left = np.mean(lm_eye_left, axis=0)
+        eye_right = np.mean(lm_eye_right, axis=0)
+        eye_avg = (eye_left + eye_right) * 0.5
+        eye_to_eye = eye_right - eye_left
+        mouth_left = lm_mouth_outer[0]
+        mouth_right = lm_mouth_outer[6]
+        mouth_avg = (mouth_left + mouth_right) * 0.5
+        eye_to_mouth = mouth_avg - eye_avg
+
+        # Choose oriented crop rectangle.
+        x = eye_to_eye - np.flipud(eye_to_mouth) * [-1, 1]
+        x /= np.hypot(*x)
+        x *= max(np.hypot(*eye_to_eye) * 2.0, np.hypot(*eye_to_mouth) * 1.8)
+        y = np.flipud(x) * [-1, 1]
+        c = eye_avg + eye_to_mouth * 0.1
+        quad = np.stack([c - x - y, c - x + y, c + x + y, c + x - y])
+        qsize = np.hypot(*x) * 2
+
+        # read image
+        if is_filepath:
+            img = PIL.Image.open(data)
+        else:
+            img = images[num_img]
+
+        output_size = 1024
+        # output_size = 256
+        transform_size = 4096
+        enable_padding = True
+
+        # Shrink.
+        shrink = int(np.floor(qsize / output_size * 0.5))
+        if shrink > 1:
+            rsize = (int(np.rint(float(img.size[0]) / shrink)), int(np.rint(float(img.size[1]) / shrink)))
+            img = img.resize(rsize, PIL.Image.ANTIALIAS)
+            quad /= shrink
+            qsize /= shrink
+
+        # Crop.
+        border = max(int(np.rint(qsize * 0.1)), 3)
+        crop = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), int(np.ceil(max(quad[:, 0]))),
+                int(np.ceil(max(quad[:, 1]))))
+        crop = (max(crop[0] - border, 0), max(crop[1] - border, 0), min(crop[2] + border, img.size[0]),
+                min(crop[3] + border, img.size[1]))
+        if crop[2] - crop[0] < img.size[0] or crop[3] - crop[1] < img.size[1]:
+            img = img.crop(crop)
+            quad -= crop[0:2]
+
+        # Pad.
+        pad = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), int(np.ceil(max(quad[:, 0]))),
+               int(np.ceil(max(quad[:, 1]))))
+        pad = (max(-pad[0] + border, 0), max(-pad[1] + border, 0), max(pad[2] - img.size[0] + border, 0),
+               max(pad[3] - img.size[1] + border, 0))
+        if enable_padding and max(pad) > border - 4:
+            pad = np.maximum(pad, int(np.rint(qsize * 0.3)))
+            img = np.pad(np.float32(img), ((pad[1], pad[3]), (pad[0], pad[2]), (0, 0)), 'reflect')
+            h, w, _ = img.shape
+            y, x, _ = np.ogrid[:h, :w, :1]
+            mask = np.maximum(1.0 - np.minimum(np.float32(x) / pad[0], np.float32(w - 1 - x) / pad[2]),
+                              1.0 - np.minimum(np.float32(y) / pad[1], np.float32(h - 1 - y) / pad[3]))
+            blur = qsize * 0.02
+            img += (scipy.ndimage.gaussian_filter(img, [blur, blur, 0]) - img) * np.clip(mask * 3.0 + 1.0, 0.0, 1.0)
+            img += (np.median(img, axis=(0, 1)) - img) * np.clip(mask, 0.0, 1.0)
+            img = PIL.Image.fromarray(np.uint8(np.clip(np.rint(img), 0, 255)), 'RGB')
+            quad += pad[:2]
+
+        # Transform.
+        img = img.transform((transform_size, transform_size), PIL.Image.QUAD, (quad + 0.5).flatten(),
+                            PIL.Image.BILINEAR)
+        if output_size < transform_size:
+            img = img.resize((output_size, output_size), PIL.Image.LANCZOS)
+
+        # Save aligned image.
+        imgs.append(img)
+
+    if return_tensors:
+        transform = T.ToTensor()
+        tensors = [transform(img).clamp(0, 1) for img in imgs]
+        return tensors
+    return imgs