eval_onnx.py

#!/usr/bin/env python3
# -*- coding:utf-8 -*-

import io
import sys
import cv2
import json
import time
import pathlib
import argparse
import tempfile
import itertools
import contextlib
import torch
import torchvision
import numpy as np
import onnxruntime as ort
from tqdm import tqdm
from loguru import logger
from tabulate import tabulate
from collections import defaultdict
from pycocotools.cocoeval import COCOeval

CURRENT_DIR = pathlib.Path(__file__).parent
sys.path.append(str(CURRENT_DIR))

from coco import COCO_CLASSES


class COCOEvaluator:
    """
    COCO AP Evaluation class.  All the data in the val2017 dataset are processed
    and evaluated by COCO API.
    """

    def __init__(
        self,
        dataloader,
        img_size: int,
        confthre: float,
        nmsthre: float,
        num_classes: int,
        testdev: bool = False,
        per_class_AP: bool = False,
        per_class_AR: bool = False,
    ):
        """
        Args:
            dataloader (Dataloader): evaluate dataloader.
            img_size: image size after preprocess. images are resized
                to squares whose shape is (img_size, img_size).
            confthre: confidence threshold ranging from 0 to 1, which
                is defined in the config file.
            nmsthre: IoU threshold of non-max supression ranging from 0 to 1.
            num_classes: number of all classes of interest.
            testdev: whether run on the testdev set of COCO.
            per_class_AP: Show per class AP during evalution or not. Default to False.
            per_class_AR: Show per class AR during evalution or not. Default to False.
        """
        self.dataloader = dataloader
        self.img_size = img_size
        self.confthre = confthre
        self.nmsthre = nmsthre
        self.num_classes = num_classes
        self.testdev = testdev
        self.per_class_AP = per_class_AP
        self.per_class_AR = per_class_AR

    def evaluate(self, ort_sess, return_outputs=False):
        """
        COCO average precision (AP) Evaluation. Iterate inference on the test dataset
        and the results are evaluated by COCO API.

        NOTE: This function will change training mode to False, please save states if needed.

        Args:
            ort_sess (onnxruntime.InferenceSession): onnxruntime session to evaluate.
            return_outputs (bool): flag indicates whether return image-wise result or not

        Returns:
            eval_results (tuple): summary of metrics for evaluation
            output_data (defaultdict): image-wise result
        """
        data_list = []
        output_data = defaultdict()
        inference_time = 0
        nms_time = 0
        n_samples = max(len(self.dataloader) - 1, 1)
        input_name = ort_sess.get_inputs()[0].name
        for cur_iter, (imgs, _, info_imgs, ids) in enumerate(tqdm(self.dataloader)):
            # with torch.no_grad():
            # skip the last iters since batchsize might be not enough for batch inference
            is_time_record = cur_iter < len(self.dataloader) - 1
            if is_time_record:
                start = time.time()
            outputs = ort_sess.run(None, {input_name: imgs.numpy()})
            outputs = [torch.Tensor(out) for out in outputs]
            outputs = head_postprocess(outputs)
            if is_time_record:
                infer_end = time.time()
                inference_time += infer_end - start
            outputs = postprocess(outputs, self.num_classes, self.confthre, self.nmsthre)
            if is_time_record:
                nms_end = time.time()
                nms_time += nms_end - infer_end
            data_list_elem, image_wise_data = self.convert_to_coco_format(
                outputs, info_imgs, ids, return_outputs=True)
            data_list.extend(data_list_elem)
            output_data.update(image_wise_data)
        statistics = [inference_time, nms_time, n_samples]
        eval_results = self.evaluate_prediction(data_list, statistics)
        if return_outputs:
            return eval_results, output_data
        return eval_results

    def convert_to_coco_format(self, outputs, info_imgs, ids, return_outputs=False):
        data_list = []
        image_wise_data = defaultdict(dict)
        for (output, img_h, img_w, img_id) in zip(
            outputs, info_imgs[0], info_imgs[1], ids
        ):
            if output is None:
                continue
            output = output.cpu()
            bboxes = output[:, 0:4]
            # preprocessing: resize
            scale = min(
                self.img_size[0] / float(img_h), self.img_size[1] / float(img_w)
            )
            bboxes /= scale
            cls = output[:, 6]
            scores = output[:, 4] * output[:, 5]
            image_wise_data.update({
                int(img_id): {
                    "bboxes": [box.numpy().tolist() for box in bboxes],
                    "scores": [score.numpy().item() for score in scores],
                    "categories": [
                        self.dataloader.dataset.class_ids[int(cls[ind])]
                        for ind in range(bboxes.shape[0])
                    ],
                }
            })
            bboxes = xyxy2xywh(bboxes)
            for ind in range(bboxes.shape[0]):
                label = self.dataloader.dataset.class_ids[int(cls[ind])]
                pred_data = {
                    "image_id": int(img_id),
                    "category_id": label,
                    "bbox": bboxes[ind].numpy().tolist(),
                    "score": scores[ind].numpy().item(),
                    "segmentation": [],
                }  # COCO json format
                data_list.append(pred_data)
        if return_outputs:
            return data_list, image_wise_data
        return data_list

    def evaluate_prediction(self, data_dict, statistics):
        # if not is_main_process():
        #     return 0, 0, None
        logger.info("Evaluate in main process...")
        annType = ["segm", "bbox", "keypoints"]
        inference_time = statistics[0]
        nms_time = statistics[1]
        n_samples = statistics[2]
        a_infer_time = 1000 * inference_time / (n_samples * self.dataloader.batch_size)
        a_nms_time = 1000 * nms_time / (n_samples * self.dataloader.batch_size)
        time_info = ", ".join(
            [
                "Average {} time: {:.2f} ms".format(k, v)
                for k, v in zip(
                    ["forward", "NMS", "inference"],
                    [a_infer_time, a_nms_time, (a_infer_time + a_nms_time)],
                )
            ]
        )
        info = time_info + "\n"
        # Evaluate the Dt (detection) json comparing with the ground truth
        if len(data_dict) > 0:
            cocoGt = self.dataloader.dataset.coco
            if self.testdev:
                json.dump(data_dict, open("./yolox_testdev_2017.json", "w"))
                cocoDt = cocoGt.loadRes("./yolox_testdev_2017.json")
            else:
                _, tmp = tempfile.mkstemp()
                json.dump(data_dict, open(tmp, "w"))
                cocoDt = cocoGt.loadRes(tmp)
            logger.info("Use standard COCOeval.")
            cocoEval = COCOeval(cocoGt, cocoDt, annType[1])
            cocoEval.evaluate()
            cocoEval.accumulate()
            redirect_string = io.StringIO()
            with contextlib.redirect_stdout(redirect_string):
                cocoEval.summarize()
            info += redirect_string.getvalue()
            cat_ids = list(cocoGt.cats.keys())
            cat_names = [cocoGt.cats[catId]['name'] for catId in sorted(cat_ids)]
            if self.per_class_AP:
                AP_table = per_class_AP_table(cocoEval, class_names=cat_names)
                info += "per class AP:\n" + AP_table + "\n"
            if self.per_class_AR:
                AR_table = per_class_AR_table(cocoEval, class_names=cat_names)
                info += "per class AR:\n" + AR_table + "\n"
            return cocoEval.stats[0], cocoEval.stats[1], info
        else:
            return 0, 0, info


class ValTransform:
    """
    Defines the transformations that should be applied to test PIL image
    for input into the network
    """

    def __init__(self, swap=(2, 0, 1), legacy=False):
        self.swap = swap
        self.legacy = legacy

    # assume input is cv2 img for now
    def __call__(self, img, res, input_size):
        img, _ = preproc(img, input_size, self.swap)
        if self.legacy:
            img = img[::-1, :, :].copy()
            img /= 255.0
            img -= np.array([0.485, 0.456, 0.406]).reshape(3, 1, 1)
            img /= np.array([0.229, 0.224, 0.225]).reshape(3, 1, 1)
        return img, np.zeros((1, 5))


def preproc(img, input_size, swap=(2, 0, 1)):
    """Preprocess function for preparing input for the network"""
    if len(img.shape) == 3:
        padded_img = np.ones((input_size[0], input_size[1], 3), dtype=np.uint8) * 114
    else:
        padded_img = np.ones(input_size, dtype=np.uint8) * 114
    r = min(input_size[0] / img.shape[0], input_size[1] / img.shape[1])
    resized_img = cv2.resize(
        img,
        (int(img.shape[1] * r), int(img.shape[0] * r)),
        interpolation=cv2.INTER_LINEAR,
    ).astype(np.uint8)
    padded_img[: int(img.shape[0] * r), : int(img.shape[1] * r)] = resized_img
    padded_img = padded_img.transpose(swap)
    padded_img = np.ascontiguousarray(padded_img, dtype=np.float32)
    return padded_img, r


def postprocess(prediction, num_classes, conf_thre=0.7, nms_thre=0.45, class_agnostic=False):
    """Post-processing part after the prediction heads with NMS"""
    box_corner = prediction.new(prediction.shape)
    box_corner[:, :, 0] = prediction[:, :, 0] - prediction[:, :, 2] / 2
    box_corner[:, :, 1] = prediction[:, :, 1] - prediction[:, :, 3] / 2
    box_corner[:, :, 2] = prediction[:, :, 0] + prediction[:, :, 2] / 2
    box_corner[:, :, 3] = prediction[:, :, 1] + prediction[:, :, 3] / 2
    prediction[:, :, :4] = box_corner[:, :, :4]
    output = [None for _ in range(len(prediction))]
    for i, image_pred in enumerate(prediction):
        # If none are remaining => process next image
        if not image_pred.size(0):
            continue
        # Get score and class with the highest confidence
        class_conf, class_pred = torch.max(image_pred[:, 5: 5 + num_classes], 1, keepdim=True)
        conf_mask = (image_pred[:, 4] * class_conf.squeeze() >= conf_thre).squeeze()
        # Detections ordered as (x1, y1, x2, y2, obj_conf, class_conf, class_pred)
        detections = torch.cat((image_pred[:, :5], class_conf, class_pred.float()), 1)
        detections = detections[conf_mask]
        if not detections.size(0):
            continue
        if class_agnostic:
            nms_out_index = torchvision.ops.nms(
                detections[:, :4],
                detections[:, 4] * detections[:, 5],
                nms_thre,
            )
        else:
            nms_out_index = torchvision.ops.batched_nms(
                detections[:, :4],
                detections[:, 4] * detections[:, 5],
                detections[:, 6],
                nms_thre,
            )
        detections = detections[nms_out_index]
        if output[i] is None:
            output[i] = detections
        else:
            output[i] = torch.cat((output[i], detections))
    return output


def head_postprocess(outputs, strides=[8, 16, 32]):
    """Decode outputs from predictions of the detection heads"""
    hw = [x.shape[-2:] for x in outputs]
    # [batch, n_anchors_all, 85]
    outputs = torch.cat([x.flatten(start_dim=2) for x in outputs], dim=2).permute(0, 2, 1)
    outputs[..., 4:] = outputs[..., 4:].sigmoid()
    return decode_outputs(outputs, outputs[0].type(), hw, strides)


def decode_outputs(outputs, dtype, ori_hw, ori_strides):
    grids = []
    strides = []
    for (hsize, wsize), stride in zip(ori_hw, ori_strides):
        yv, xv = meshgrid([torch.arange(hsize), torch.arange(wsize)])
        grid = torch.stack((xv, yv), 2).view(1, -1, 2)
        grids.append(grid)
        shape = grid.shape[:2]
        strides.append(torch.full((*shape, 1), stride))
    grids = torch.cat(grids, dim=1).type(dtype)
    strides = torch.cat(strides, dim=1).type(dtype)
    outputs[..., :2] = (outputs[..., :2] + grids) * strides
    outputs[..., 2:4] = torch.exp(outputs[..., 2:4]) * strides
    return outputs


def xyxy2xywh(bboxes):
    bboxes[:, 2] = bboxes[:, 2] - bboxes[:, 0]
    bboxes[:, 3] = bboxes[:, 3] - bboxes[:, 1]
    return bboxes


def meshgrid(*tensors):
    _TORCH_VER = [int(x) for x in torch.__version__.split(".")[:2]]
    if _TORCH_VER >= [1, 10]:
        return torch.meshgrid(*tensors, indexing="ij")
    else:
        return torch.meshgrid(*tensors)


def per_class_AR_table(coco_eval, class_names=COCO_CLASSES, headers=["class", "AR"], colums=6):
    """Format the recall of each class"""
    per_class_AR = {}
    recalls = coco_eval.eval["recall"]
    # dimension of recalls: [TxKxAxM]
    # recall has dims (iou, cls, area range, max dets)
    assert len(class_names) == recalls.shape[1]
    for idx, name in enumerate(class_names):
        recall = recalls[:, idx, 0, -1]
        recall = recall[recall > -1]
        ar = np.mean(recall) if recall.size else float("nan")
        per_class_AR[name] = float(ar * 100)
    num_cols = min(colums, len(per_class_AR) * len(headers))
    result_pair = [x for pair in per_class_AR.items() for x in pair]
    row_pair = itertools.zip_longest(*[result_pair[i::num_cols] for i in range(num_cols)])
    table_headers = headers * (num_cols // len(headers))
    table = tabulate(
        row_pair, tablefmt="pipe", floatfmt=".3f", headers=table_headers, numalign="left",
    )
    return table


def per_class_AP_table(coco_eval, class_names=COCO_CLASSES, headers=["class", "AP"], colums=6):
    """Format the precision of each class"""
    per_class_AP = {}
    precisions = coco_eval.eval["precision"]
    # dimension of precisions: [TxRxKxAxM]
    # precision has dims (iou, recall, cls, area range, max dets)
    assert len(class_names) == precisions.shape[2]
    for idx, name in enumerate(class_names):
        # area range index 0: all area ranges
        # max dets index -1: typically 100 per image
        precision = precisions[:, :, idx, 0, -1]
        precision = precision[precision > -1]
        ap = np.mean(precision) if precision.size else float("nan")
        per_class_AP[name] = float(ap * 100)
    num_cols = min(colums, len(per_class_AP) * len(headers))
    result_pair = [x for pair in per_class_AP.items() for x in pair]
    row_pair = itertools.zip_longest(*[result_pair[i::num_cols] for i in range(num_cols)])
    table_headers = headers * (num_cols // len(headers))
    table = tabulate(
        row_pair, tablefmt="pipe", floatfmt=".3f", headers=table_headers, numalign="left",
    )
    return table


def get_eval_loader(batch_size, test_size=(640, 640), data_dir='data/COCO', data_num_workers=0, testdev=False, legacy=False):
    from coco import COCODataset
    valdataset = COCODataset(
        data_dir=data_dir,
        json_file='instances_val2017.json' if not testdev else 'instances_test2017.json',
        name="val2017" if not testdev else "test2017",
        img_size=test_size,
        preproc=ValTransform(legacy=legacy),
    )
    sampler = torch.utils.data.SequentialSampler(valdataset)
    dataloader_kwargs = {
        "num_workers": data_num_workers,
        "pin_memory": True,
        "sampler": sampler,
        "batch_size": batch_size
    }
    val_loader = torch.utils.data.DataLoader(valdataset, **dataloader_kwargs)
    return val_loader


def make_parser():
    parser = argparse.ArgumentParser("onnxruntime inference sample")
    parser.add_argument(
        "-m",
        "--model",
        type=str,
        default="yolox-s-int8.onnx",
        help="Input your onnx model.",
    )
    parser.add_argument(
        "-b",
        "--batch_size",
        type=int,
        default=1,
        help="Batch size for inference..",
    )
    parser.add_argument(
        "--input_shape",
        type=str,
        default="640,640",
        help="Specify an input shape for inference.",
    )
    parser.add_argument(
        "--ipu",
        action="store_true",
        help="Use IPU for inference.",
    )
    parser.add_argument(
        "--provider_config",
        type=str,
        default="vaip_config.json",
        help="Path of the config file for setting provider_options.",
    )
    return parser


if __name__ == '__main__':
    args = make_parser().parse_args()
    input_shape = tuple(map(int, args.input_shape.split(',')))
    if args.ipu:
        providers = ["VitisAIExecutionProvider"]
        provider_options = [{"config_file": args.provider_config}]
    else:
        providers = ['CUDAExecutionProvider', 'CPUExecutionProvider']
        provider_options = None
    session = ort.InferenceSession(args.model, providers=providers, provider_options=provider_options)
    val_loader = get_eval_loader(args.batch_size)
    evaluator = COCOEvaluator(dataloader=val_loader, img_size=input_shape, confthre=0.01, nmsthre=0.65, num_classes=80, testdev=False)
    *_, summary = evaluator.evaluate(session)
    logger.info("\n" + summary)