train.py

#
# Copyright (C) 2023, Inria
# GRAPHDECO research group, https://team.inria.fr/graphdeco
# All rights reserved.
#
# This software is free for non-commercial, research and evaluation use 
# under the terms of the LICENSE.md file.
#
# For inquiries contact  george.drettakis@inria.fr
#
import os
os.environ["CUDA_LAUNCH_BLOCKING"] = "1"

import torch
from random import randint
from utils.loss_utils import l1_loss, ssim, compute_depth_loss, compute_rank_loss, compute_continue_loss
from gaussian_renderer import render, network_gui, AppearanceOptimizer
import sys
from scene import Scene, GaussianModel
from utils.general_utils import safe_state
import uuid
from tqdm import tqdm
from utils.image_utils import psnr
from argparse import ArgumentParser, Namespace
from arguments import ModelParams, PipelineParams, OptimizationParams
try:
    from torch.utils.tensorboard import SummaryWriter
    TENSORBOARD_FOUND = True
except ImportError:
    TENSORBOARD_FOUND = False


def training(dataset, opt, pipe, testing_iterations, saving_iterations, checkpoint_iterations, checkpoint, debug_from, depth_loss_choice):
    first_iter = 0
    tb_writer = prepare_output_and_logger(dataset)
    gaussians = GaussianModel(dataset.sh_degree) # 首先实例化3d高斯
    scene = Scene(dataset, gaussians) # 这一步根据读取的数据去给3d高斯的属性进行初始化
    ## 实例化相机姿态优化类
    # cameraoptimizer = CameraOptimizer(len(scene.getTrainCameras()))

    ## 实例化外观嵌入类
    if dataset.able_appearance_embedding:
        print('Using Appearance Optimizer')
        appearanceoptimizer = AppearanceOptimizer(len(scene.getTrainCameras()))
    else:
        print('Appearance Optimizer Close')
        appearanceoptimizer = None

    # 检测深度图是否准备就绪
    if scene.getTrainCameras()[0].depth is not None and dataset.using_depth:
        print("Depth map Ready！")
    else:
        assert dataset.using_depth == False, "depth map is not exist, so using depth must close"

    gaussians.training_setup(opt) # 训练前的准备工作
    if checkpoint:
        (model_params, first_iter) = torch.load(checkpoint)
        gaussians.restore(model_params, opt)

    bg_color = [1, 1, 1] if dataset.white_background else [0, 0, 0]
    background = torch.tensor(bg_color, dtype=torch.float32, device="cuda")

    iter_start = torch.cuda.Event(enable_timing = True)
    iter_end = torch.cuda.Event(enable_timing = True)

    viewpoint_stack = None
    ema_loss_for_log = 0.0
    progress_bar = tqdm(range(first_iter, opt.iterations), desc="Training progress")
    first_iter += 1
    for iteration in range(first_iter, opt.iterations + 1):        
        if network_gui.conn == None:
            network_gui.try_connect()
        while network_gui.conn != None:
            try:
                net_image_bytes = None
                custom_cam, do_training, pipe.convert_SHs_python, pipe.compute_cov3D_python, keep_alive, scaling_modifer = network_gui.receive()
                if custom_cam != None:
                    net_image = render(custom_cam, gaussians, pipe, background, scaling_modifer)["render"]
                    net_image_bytes = memoryview((torch.clamp(net_image, min=0, max=1.0) * 255).byte().permute(1, 2, 0).contiguous().cpu().numpy())
                network_gui.send(net_image_bytes, dataset.source_path)
                if do_training and ((iteration < int(opt.iterations)) or not keep_alive):
                    break
            except Exception as e:
                network_gui.conn = None

        iter_start.record()

        gaussians.update_learning_rate(iteration) # 使用学习率衰减策略用于高斯

        # Every 1000 its we increase the levels of SH up to a maximum degree
        if iteration % 1000 == 0:
            gaussians.oneupSHdegree()

        # Pick a random Camera
        if not viewpoint_stack:
            viewpoint_stack = scene.getTrainCameras().copy()
        viewpoint_cam = viewpoint_stack.pop(randint(0, len(viewpoint_stack)-1)) # 随机从训练集上选择一个视点

        if dataset.able_appearance_embedding:
            # appearance embedding
            rgb_factors = appearanceoptimizer(viewpoint_cam)
        else:
            rgb_factors = None
        # Render
        if (iteration - 1) == debug_from:
            pipe.debug = True
        render_pkg = render(viewpoint_cam, gaussians, pipe, background, rgb_factors=rgb_factors)
        image, viewspace_point_tensor, visibility_filter, radii, depth = render_pkg["render"], render_pkg["viewspace_points"], render_pkg["visibility_filter"], render_pkg["radii"], render_pkg["depth"]

        # Loss
        gt_image = viewpoint_cam.original_image.cuda()
        Ll1 = l1_loss(image, gt_image)
        loss = (1.0 - opt.lambda_dssim) * Ll1 + opt.lambda_dssim * (1.0 - ssim(image, gt_image))
        # add depth loss
        if viewpoint_cam.depth is not None and dataset.using_depth:
            assert args.depth_loss_choice in ["localrf", "rank_loss", "continue_loss",
                                              "hybrid_loss", "L1_loss"], "loss choice error!"
            if depth_loss_choice == 'localrf':
                gt_depth = viewpoint_cam.depth.cuda()
                depth_loss = compute_depth_loss(1 / depth.clamp(1e-6), gt_depth, opt.lambda_depth)
                loss = loss + depth_loss
            elif depth_loss_choice == 'rank_loss':
                gt_depth = viewpoint_cam.depth.cuda()
                depth_loss = compute_rank_loss(1 / depth.clamp(1e-6), gt_depth, opt.lambda_rank_depth)
                loss = loss + depth_loss
            elif depth_loss_choice == 'continue_loss':
                gt_depth = viewpoint_cam.depth.cuda()
                depth_loss = compute_continue_loss(1 / depth.clamp(1e-6), gt_depth, opt.lambda_continue_depth)
                loss = loss + depth_loss
            elif depth_loss_choice == 'hybrid_loss':
                gt_depth = viewpoint_cam.depth.cuda()
                depth_loss_continue = compute_continue_loss(1 / depth.clamp(1e-6), gt_depth, opt.lambda_continue_depth)
                depth_loss_rank = compute_rank_loss(1 / depth.clamp(1e-6), gt_depth, opt.lambda_rank_depth)
                depth_loss = depth_loss_continue + depth_loss_rank
                loss = loss + depth_loss
            elif depth_loss_choice == 'L1_loss':
                 # 应该处理gt——depth而不是pred——depth
                gt_depth = viewpoint_cam.depth.cuda()
                gt_depth = gt_depth / gt_depth.max()
                depth_loss = l1_loss(1 / depth.clamp(1e-6), gt_depth) * opt.lambda_depth
                loss = loss + depth_loss

        loss.backward()
        iter_end.record()

        with torch.no_grad():
            # Progress bar
            ema_loss_for_log = 0.4 * loss.item() + 0.6 * ema_loss_for_log
            if iteration % 10 == 0:
                progress_bar.set_postfix({"Loss": f"{ema_loss_for_log:.{7}f}",
                                          "totol points": f"{scene.gaussians.get_xyz.shape[0]}"
                                          })
                progress_bar.update(10)
            if iteration == opt.iterations:
                progress_bar.close()

            # Log and save
            if depth_loss_choice is not None:
                training_report_add_depth(tb_writer, iteration, Ll1, depth_loss, loss, l1_loss, iter_start.elapsed_time(iter_end),
                                testing_iterations, scene, render, (pipe, background), depth_loss_choice)
            else:
                training_report(tb_writer, iteration, Ll1, loss, l1_loss, iter_start.elapsed_time(iter_end), testing_iterations, scene, render, (pipe, background))
            if (iteration in saving_iterations):
                print("\n[ITER {}] Saving Gaussians".format(iteration))
                scene.save(iteration)
                if dataset.able_appearance_embedding:
                    # save appearance
                    save_path = os.path.join(dataset.model_path, "point_cloud/iteration_{}".format(iteration))
                    appearanceoptimizer.save_appearance_embedding(os.path.join(save_path, "appearance_embedding.ckpt"))

            # Densification
            if iteration < opt.densify_until_iter: #只在前面的step进行？
                # Keep track of max radii in image-space for pruning
                gaussians.max_radii2D[visibility_filter] = torch.max(gaussians.max_radii2D[visibility_filter], radii[visibility_filter])
                gaussians.add_densification_stats(viewspace_point_tensor, visibility_filter)

                if iteration > opt.densify_from_iter and iteration % opt.densification_interval == 0:
                    size_threshold = 20 if iteration > opt.opacity_reset_interval else None
                    gaussians.densify_and_prune(opt.densify_grad_threshold, 0.005, scene.cameras_extent, size_threshold)
                
                if iteration % opt.opacity_reset_interval == 0 or (dataset.white_background and iteration == opt.densify_from_iter):
                    gaussians.reset_opacity()

            # Optimizer step
            if iteration < opt.iterations:
                gaussians.optimizer.step()
                gaussians.optimizer.zero_grad(set_to_none = True)
                # cameraoptimizer.optimizer.step()
                # cameraoptimizer.optimizer.zero_grad(set_to_none=True)
                if dataset.able_appearance_embedding:
                    appearanceoptimizer.appearance_embedding_optimizer.step()
                    appearanceoptimizer.appearance_embedding_optimizer.zero_grad(set_to_none=True)

            if (iteration in checkpoint_iterations):
                print("\n[ITER {}] Saving Checkpoint".format(iteration))
                torch.save((gaussians.capture(), iteration), scene.model_path + "/chkpnt" + str(iteration) + ".pth")

def prepare_output_and_logger(args):    
    if not args.model_path:
        if os.getenv('OAR_JOB_ID'):
            unique_str=os.getenv('OAR_JOB_ID')
        else:
            unique_str = str(uuid.uuid4())
        args.model_path = os.path.join("./output/", unique_str[0:10])
        
    # Set up output folder
    print("Output folder: {}".format(args.model_path))
    os.makedirs(args.model_path, exist_ok = True)
    with open(os.path.join(args.model_path, "cfg_args"), 'w') as cfg_log_f:
        cfg_log_f.write(str(Namespace(**vars(args))))

    # Create Tensorboard writer
    tb_writer = None
    if TENSORBOARD_FOUND:
        tb_writer = SummaryWriter(args.model_path)
    else:
        print("Tensorboard not available: not logging progress")
    return tb_writer

def training_report(tb_writer, iteration, Ll1, loss, l1_loss, elapsed, testing_iterations, scene : Scene, renderFunc, renderArgs):
    if tb_writer:
        tb_writer.add_scalar('train_loss_patches/l1_loss', Ll1.item(), iteration)
        tb_writer.add_scalar('train_loss_patches/total_loss', loss.item(), iteration)
        tb_writer.add_scalar('iter_time', elapsed, iteration)

    # Report test and samples of training set
    if iteration in testing_iterations:
        torch.cuda.empty_cache()
        validation_configs = ({'name': 'test', 'cameras' : scene.getTestCameras()}, 
                              {'name': 'train', 'cameras' : [scene.getTrainCameras()[idx % len(scene.getTrainCameras())] for idx in range(5, 30, 5)]})

        for config in validation_configs:
            if config['cameras'] and len(config['cameras']) > 0:
                l1_test = 0.0
                psnr_test = 0.0
                for idx, viewpoint in enumerate(config['cameras']):
                    image = torch.clamp(renderFunc(viewpoint, scene.gaussians, *renderArgs)["render"], 0.0, 1.0)
                    gt_image = torch.clamp(viewpoint.original_image.to("cuda"), 0.0, 1.0)
                    if tb_writer and (idx < 5):
                        tb_writer.add_images(config['name'] + "_view_{}/render".format(viewpoint.image_name), image[None], global_step=iteration)
                        if iteration == testing_iterations[0]:
                            tb_writer.add_images(config['name'] + "_view_{}/ground_truth".format(viewpoint.image_name), gt_image[None], global_step=iteration)
                    l1_test += l1_loss(image, gt_image).mean().double()
                    psnr_test += psnr(image, gt_image).mean().double()
                psnr_test /= len(config['cameras'])
                l1_test /= len(config['cameras'])          
                print("\n[ITER {}] Evaluating {}: L1 {} PSNR {}".format(iteration, config['name'], l1_test, psnr_test))
                if tb_writer:
                    tb_writer.add_scalar(config['name'] + '/loss_viewpoint - l1_loss', l1_test, iteration)
                    tb_writer.add_scalar(config['name'] + '/loss_viewpoint - psnr', psnr_test, iteration)

        if tb_writer:
            tb_writer.add_histogram("scene/opacity_histogram", scene.gaussians.get_opacity, iteration)
            tb_writer.add_scalar('total_points', scene.gaussians.get_xyz.shape[0], iteration)
        torch.cuda.empty_cache()

def training_report_add_depth(tb_writer, iteration, Ll1, depth_loss, loss, l1_loss, elapsed, testing_iterations, scene : Scene, renderFunc, renderArgs, depth_loss_choice):
    if tb_writer:
        tb_writer.add_scalar('train_loss_patches/l1_loss', Ll1.item(), iteration)
        tb_writer.add_scalar(f'train_loss_patches/{depth_loss_choice}', depth_loss.item(), iteration)
        tb_writer.add_scalar('train_loss_patches/total_loss', loss.item(), iteration)
        tb_writer.add_scalar('iter_time', elapsed, iteration)

    # Report test and samples of training set
    if iteration in testing_iterations:
        torch.cuda.empty_cache()
        validation_configs = ({'name': 'test', 'cameras' : scene.getTestCameras()},
                              {'name': 'train', 'cameras' : [scene.getTrainCameras()[idx % len(scene.getTrainCameras())] for idx in range(5, 30, 5)]})

        for config in validation_configs:
            if config['cameras'] and len(config['cameras']) > 0:
                l1_test = 0.0
                psnr_test = 0.0
                for idx, viewpoint in enumerate(config['cameras']):
                    image = torch.clamp(renderFunc(viewpoint, scene.gaussians, *renderArgs)["render"], 0.0, 1.0)
                    gt_image = torch.clamp(viewpoint.original_image.to("cuda"), 0.0, 1.0)
                    if tb_writer and (idx < 5):
                        tb_writer.add_images(config['name'] + "_view_{}/render".format(viewpoint.image_name), image[None], global_step=iteration)
                        if iteration == testing_iterations[0]:
                            tb_writer.add_images(config['name'] + "_view_{}/ground_truth".format(viewpoint.image_name), gt_image[None], global_step=iteration)
                    l1_test += l1_loss(image, gt_image).mean().double()
                    psnr_test += psnr(image, gt_image).mean().double()
                psnr_test /= len(config['cameras'])
                l1_test /= len(config['cameras'])
                print("\n[ITER {}] Evaluating {}: L1 {} PSNR {}".format(iteration, config['name'], l1_test, psnr_test))
                if tb_writer:
                    tb_writer.add_scalar(config['name'] + '/loss_viewpoint - l1_loss', l1_test, iteration)
                    tb_writer.add_scalar(config['name'] + '/loss_viewpoint - psnr', psnr_test, iteration)

        if tb_writer:
            tb_writer.add_histogram("scene/opacity_histogram", scene.gaussians.get_opacity, iteration)
            tb_writer.add_scalar('total_points', scene.gaussians.get_xyz.shape[0], iteration)
        torch.cuda.empty_cache()

if __name__ == "__main__":
    # Set up command line argument parser
    parser = ArgumentParser(description="Training script parameters")
    lp = ModelParams(parser)
    op = OptimizationParams(parser)
    pp = PipelineParams(parser)
    parser.add_argument('--ip', type=str, default="127.0.0.1")
    parser.add_argument('--port', type=int, default=6007)
    parser.add_argument('--debug_from', type=int, default=-1)
    parser.add_argument('--detect_anomaly', action='store_true', default=False)
    parser.add_argument("--test_iterations", nargs="+", type=int, default=[3_000, 7_000, 30_000, 50_000, 100_000])
    parser.add_argument("--save_iterations", nargs="+", type=int, default=[3_000, 7_000, 30_000, 50_000, 100_000])
    parser.add_argument("--quiet", action="store_true")
    parser.add_argument("--checkpoint_iterations", nargs="+", type=int, default=[])
    parser.add_argument("--start_checkpoint", type=str, default=None)
    parser.add_argument("--depth_loss_choice", type=str, default=None)
    args = parser.parse_args(sys.argv[1:])
    args.save_iterations.append(args.iterations)
    
    print("Optimizing " + args.model_path)
    if lp.extract(args).using_depth:
        print(f'using depth supervision {args.depth_loss_choice}')
        assert args.depth_loss_choice in ["localrf", "rank_loss", "continue_loss",
                                          "hybrid_loss", "L1_loss"], "loss choice error!"
    else:
        print('depth supervision close')
    # Initialize system state (RNG)
    safe_state(args.quiet)

    # Start GUI server, configure and run training
    network_gui.init(args.ip, args.port)
    torch.autograd.set_detect_anomaly(args.detect_anomaly)
    training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from, args.depth_loss_choice)

    # All done
    print("\nTraining complete.")