import math
import sys
import time
import torch
import torchvision.models.detection.mask_rcnn
from .coco_eval import CocoEvaluator
from .coco_utils import get_coco_api_from_dataset
from .utils import MetricLogger, SmoothedValue, reduce_dict
[docs]
def train_one_epoch(
model, optimizer, data_loader, device, epoch, print_freq, scaler=None
):
model.train()
metric_logger = MetricLogger(delimiter=" ")
metric_logger.add_meter(
"lr", SmoothedValue(window_size=1, fmt="{value:.6f}")
)
header = f"Epoch: [{epoch}]"
lr_scheduler = None
if epoch == 0:
warmup_factor = 1.0 / 1000
warmup_iters = min(1000, len(data_loader) - 1)
lr_scheduler = torch.optim.lr_scheduler.LinearLR(
optimizer, start_factor=warmup_factor, total_iters=warmup_iters
)
for images, targets in metric_logger.log_every(
data_loader, print_freq, header
):
input_images = [image.float().to(device) for image in images]
train_targets = [
{
k: v.to(torch.int64).to(device)
if isinstance(v, torch.Tensor)
else v
for k, v in t.items()
}
for t in targets
]
with torch.amp.autocast(str(device), enabled=scaler is not None):
loss_dict = model(input_images, train_targets)
losses = sum(loss for loss in loss_dict.values())
# reduce losses over all GPUs for logging purposes
loss_dict_reduced = reduce_dict(loss_dict)
losses_reduced = sum(loss for loss in loss_dict_reduced.values())
loss_value = losses_reduced.item()
if not math.isfinite(loss_value):
print(f"Loss is {loss_value}, stopping training")
print(loss_dict_reduced)
sys.exit(1)
optimizer.zero_grad()
if scaler is not None:
scaler.scale(losses).backward()
scaler.step(optimizer)
scaler.update()
else:
losses.backward()
optimizer.step()
if lr_scheduler is not None:
lr_scheduler.step()
metric_logger.update(loss=losses_reduced, **loss_dict_reduced)
metric_logger.update(lr=optimizer.param_groups[0]["lr"])
return metric_logger
def _get_iou_types(model):
model_without_ddp = model
if isinstance(model, torch.nn.parallel.DistributedDataParallel):
model_without_ddp = model.module
iou_types = ["bbox"]
if isinstance(model_without_ddp, torchvision.models.detection.MaskRCNN):
iou_types.append("segm")
if isinstance(model_without_ddp, torchvision.models.detection.KeypointRCNN):
iou_types.append("keypoints")
return iou_types
[docs]
@torch.inference_mode()
def evaluate(model, data_loader, device):
n_threads = torch.get_num_threads()
# FIXME remove this and make paste_masks_in_image run on the GPU
torch.set_num_threads(1)
cpu_device = torch.device("cpu")
model.eval()
metric_logger = MetricLogger(delimiter=" ")
header = "Test:"
coco = get_coco_api_from_dataset(data_loader.dataset)
iou_types = _get_iou_types(model)
coco_evaluator = CocoEvaluator(coco, iou_types)
for imgs, targets in metric_logger.log_every(data_loader, 100, header):
images = [img.to(device).float() for img in imgs]
if torch.cuda.is_available():
torch.cuda.synchronize()
model_time = time.time()
outputs = model(images)
outputs = [{k: v.to(cpu_device) for k, v in t.items()} for t in outputs]
model_time = time.time() - model_time
res = {
target["image_id"]: output
for target, output in zip(targets, outputs)
}
evaluator_time = time.time()
coco_evaluator.update(res)
evaluator_time = time.time() - evaluator_time
metric_logger.update(
model_time=model_time, evaluator_time=evaluator_time
)
# gather the stats from all processes
metric_logger.synchronize_between_processes()
print("Averaged stats:", metric_logger)
coco_evaluator.synchronize_between_processes()
# accumulate predictions from all images
coco_evaluator.accumulate()
coco_evaluator.summarize()
torch.set_num_threads(n_threads)
return coco_evaluator