model.py

from __future__ import annotations

import os
import pathlib
import shlex
import subprocess
import sys

if os.getenv('SYSTEM') == 'spaces':
    import mim

    mim.uninstall('mmcv-full', confirm_yes=True)
    mim.install('mmcv-full==1.5.0', is_yes=True)

    subprocess.run(shlex.split('pip uninstall -y opencv-python'))
    subprocess.run(shlex.split('pip uninstall -y opencv-python-headless'))
    subprocess.run(shlex.split('pip install opencv-python-headless==4.8.0.74'))

    with open('patch') as f:
        subprocess.run(shlex.split('patch -p1'), cwd='CBNetV2', stdin=f)
    subprocess.run('mv palette.py CBNetV2/mmdet/core/visualization/'.split())

import numpy as np
import torch
import torch.nn as nn

app_dir = pathlib.Path(__file__).parent
submodule_dir = app_dir / 'CBNetV2/'
sys.path.insert(0, submodule_dir.as_posix())

from mmdet.apis import inference_detector, init_detector


class Model:
    def __init__(self):
        self.device = torch.device(
            'cuda:0' if torch.cuda.is_available() else 'cpu')
        self.models = self._load_models()
        self.model_name = 'Improved HTC (DB-Swin-B)'

    def _load_models(self) -> dict[str, nn.Module]:
        model_dict = {
            'Faster R-CNN (DB-ResNet50)': {
                'config':
                'CBNetV2/configs/cbnet/faster_rcnn_cbv2d1_r50_fpn_1x_coco.py',
                'model':
                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/faster_rcnn_cbv2d1_r50_fpn_1x_coco.pth.zip',
            },
            'Mask R-CNN (DB-Swin-T)': {
                'config':
                'CBNetV2/configs/cbnet/mask_rcnn_cbv2_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.py',
                'model':
                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/mask_rcnn_cbv2_swin_tiny_patch4_window7_mstrain_480-800_adamw_3x_coco.pth.zip',
            },
            #        'Cascade Mask R-CNN (DB-Swin-S)': {
            #            'config':
            #                'CBNetV2/configs/cbnet/cascade_mask_rcnn_cbv2_swin_small_patch4_window7_mstrain_400-1400_adamw_3x_coco.py',
            #            'model':
            #                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/cascade_mask_rcnn_cbv2_swin_small_patch4_window7_mstrain_400-1400_adamw_3x_coco.pth.zip',
            #        },
            'Improved HTC (DB-Swin-B)': {
                'config':
                'CBNetV2/configs/cbnet/htc_cbv2_swin_base_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_20e_coco.py',
                'model':
                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/htc_cbv2_swin_base22k_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_20e_coco.pth.zip',
            },
            'Improved HTC (DB-Swin-L)': {
                'config':
                'CBNetV2/configs/cbnet/htc_cbv2_swin_large_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_1x_coco.py',
                'model':
                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/htc_cbv2_swin_large22k_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_1x_coco.pth.zip',
            },
            'Improved HTC (DB-Swin-L (TTA))': {
                'config':
                'CBNetV2/configs/cbnet/htc_cbv2_swin_large_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_1x_coco.py',
                'model':
                'https://github.com/CBNetwork/storage/releases/download/v1.0.0/htc_cbv2_swin_large22k_patch4_window7_mstrain_400-1400_giou_4conv1f_adamw_1x_coco.pth.zip',
            },
        }

        weight_dir = pathlib.Path('weights')
        weight_dir.mkdir(exist_ok=True)

        def _download(model_name: str, out_dir: pathlib.Path) -> None:
            import zipfile

            model_url = model_dict[model_name]['model']
            zip_name = model_url.split('/')[-1]

            out_path = out_dir / zip_name
            if out_path.exists():
                return
            torch.hub.download_url_to_file(model_url, out_path)

            with zipfile.ZipFile(out_path) as f:
                f.extractall(out_dir)

        def _get_model_path(model_name: str) -> str:
            model_url = model_dict[model_name]['model']
            model_name = model_url.split('/')[-1][:-4]
            return (weight_dir / model_name).as_posix()

        for model_name in model_dict:
            _download(model_name, weight_dir)

        models = {
            key: init_detector(dic['config'],
                               _get_model_path(key),
                               device=self.device)
            for key, dic in model_dict.items()
        }
        return models

    def set_model_name(self, name: str) -> None:
        self.model_name = name

    def detect_and_visualize(
            self, image: np.ndarray,
            score_threshold: float) -> tuple[list[np.ndarray], np.ndarray]:
        out = self.detect(image)
        vis = self.visualize_detection_results(image, out, score_threshold)
        return out, vis

    def detect(self, image: np.ndarray) -> list[np.ndarray]:
        image = image[:, :, ::-1]  # RGB -> BGR
        model = self.models[self.model_name]
        out = inference_detector(model, image)
        return out

    def visualize_detection_results(
            self,
            image: np.ndarray,
            detection_results: list[np.ndarray],
            score_threshold: float = 0.3) -> np.ndarray:
        image = image[:, :, ::-1]  # RGB -> BGR
        model = self.models[self.model_name]
        vis = model.show_result(image,
                                detection_results,
                                score_thr=score_threshold,
                                bbox_color=None,
                                text_color=(200, 200, 200),
                                mask_color=None)
        return vis[:, :, ::-1]  # BGR -> RGB