Spaces:
Runtime error
Runtime error
| import torch | |
| from torch.utils.tensorboard import SummaryWriter | |
| class Visualizer: | |
| """ | |
| Tensorboard ε―θ§εηζ§η±» | |
| """ | |
| def __init__(self, opt): | |
| """ """ | |
| self.opt = opt # cache the option | |
| self.writer = SummaryWriter(log_dir=opt.log_dir) | |
| def display_current_results(self, iters, visuals_dict): | |
| """ | |
| Display current images | |
| Parameters: | |
| ---------- | |
| visuals (OrderedDict) - - dictionary of images to display | |
| iters (int) - - the current iteration | |
| """ | |
| for label, image in visuals_dict.items(): | |
| if image.shape[0] >= 2: | |
| image = image[0:2, :, :, :] | |
| self.writer.add_images(str(label), (image * 255.0).to(torch.uint8), global_step=iters, dataformats="NCHW") | |
| def plot_current_losses(self, iters, loss_dict): | |
| """ | |
| Display losses on tensorboard | |
| Parameters: | |
| iters (int) -- current iteration | |
| losses (OrderedDict) -- training losses stored in the format of (name, torch.Tensor) pairs | |
| """ | |
| x = iters | |
| for k, v in loss_dict.items(): | |
| self.writer.add_scalar(f"Loss/{k}", v, x) | |