app.py

import gradio as gr
import tempfile
import torch
from pytorchvideo.data import make_clip_sampler
from pytorchvideo.data.clip_sampling import ClipInfoList
from pytorchvideo.data.encoded_video_pyav import EncodedVideoPyAV
from pytorchvideo.data.video import VideoPathHandler
from pytorchvideo.transforms import (
    Normalize,
    UniformTemporalSubsample, RandomShortSideScale,
)
from torchvision.transforms import (
    Compose,
    Lambda,
    Resize, RandomCrop,
)
from transformers import VideoMAEForVideoClassification, VideoMAEFeatureExtractor

from video_utils import change_video_resolution_and_fps
import numpy as np

MODEL_CKPT = "omermazig/videomae-finetuned-nba-5-class-4-batch-8000-vid-multilabel"
DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
CLIPS_FROM_SINGLE_VIDEO = 5

trained_model = VideoMAEForVideoClassification.from_pretrained(MODEL_CKPT).to(DEVICE)
image_processor = VideoMAEFeatureExtractor.from_pretrained(MODEL_CKPT)

mean = image_processor.image_mean
std = image_processor.image_std
if "shortest_edge" in image_processor.size:
    height = width = image_processor.size["shortest_edge"]
else:
    height = image_processor.size["height"]
    width = image_processor.size["width"]
resize_to = (height, width)

num_frames_to_sample = trained_model.config.num_frames
sample_rate = 4
fps = 30
clip_duration = num_frames_to_sample * sample_rate / fps

# Validation and Test datasets' transformations.
inference_transform = Compose(
                [
                    UniformTemporalSubsample(num_frames_to_sample),
                    Lambda(lambda x: x / 255.0),
                    Normalize(mean, std),
                    RandomShortSideScale(min_size=256, max_size=320),
                    RandomCrop(resize_to),
                ]
            )

num_labels = trained_model.config.num_labels
labels = [trained_model.config.id2label[i] for i in range(num_labels)]


def parse_video_to_clips(video_file):
    """A utility to parse the input videos """
    new_resolution = (320, 256)
    new_fps = 30
    acceptable_fps_violation = 5
    with tempfile.NamedTemporaryFile() as new_video:
        print(new_video.name)
        change_video_resolution_and_fps(video_file, new_video.name, new_resolution, new_fps, acceptable_fps_violation)
        video_path_handler = VideoPathHandler()
        video: EncodedVideoPyAV = video_path_handler.video_from_path(video_file)

    clip_sampler = make_clip_sampler("random_multi", clip_duration, CLIPS_FROM_SINGLE_VIDEO)
    # noinspection PyTypeChecker
    clip_info: ClipInfoList = clip_sampler(0, video.duration, {})

    video_clips_list = []
    for clip_start, clip_end in zip(clip_info.clip_start_sec, clip_info.clip_end_sec):
        video_clip = video.get_clip(clip_start, clip_end)["video"]
        video_clips_list.append(inference_transform(video_clip))

    videos_tensor = torch.stack([single_clip.permute(1, 0, 2, 3) for single_clip in video_clips_list])
    return videos_tensor.to(DEVICE)


def infer(video_file, threshold=0.5):
    videos_tensor = parse_video_to_clips(video_file)
    inputs = {"pixel_values": videos_tensor}

    # forward pass
    with torch.no_grad():
        outputs = trained_model(**inputs)
        multiple_logits = outputs.logits
        logits = multiple_logits.mean(dim=0)

    # first, apply sigmoid on logits
    sigmoid = torch.nn.Sigmoid()
    sigmoid_scores = sigmoid(torch.from_numpy(np.array(logits))).squeeze(0)
    # next, use threshold to turn them into integer predictions
    confidences = {labels[i]: float(sigmoid_scores[i]) for i in range(len(labels))}
    return confidences


gr.Interface(
    fn=infer,
    inputs=gr.Video(type="file"),
    outputs=gr.Label(num_top_classes=10),
    examples=[
        ["examples/DUNK.avi"],
        ["examples/FLOATING_JUMP_SHOT.avi"],
        ["examples/JUMP_SHOT.avi"],
        ["examples/REVERSE_LAYUP.avi"],
        ["examples/TURNAROUND_HOOK_SHOT.avi"],
    ],
    title="VideoMAE fine-tuned on nba data",
    description=(
        "Gradio demo for VideoMAE for video classification. To use it, simply upload your video or click one of the"
        " examples to load them. Read more at the links below."
    ),
    article=(
        "<div style='text-align: center;'><a href='https://huggingface.co/docs/transformers/model_doc/videomae' target='_blank'>VideoMAE</a>"
        " <center><a href='https://huggingface.co/omermazig/videomae-finetuned-nba-5-class-4-batch-8000-vid-multilabel' target='_blank'>Fine-tuned Model</a></center></div>"
    ),
    allow_flagging=False,
    allow_screenshot=False,
).launch()