并行训练代码优化

.idea/CopilotChatHistory.xml

@@ -3,6 +3,18 @@
   <component name="CopilotChatHistory">
     <option name="conversations">
       <list>
+        <Conversation>
+          <option name="createTime" value="1751441743239" />
+          <option name="id" value="0197ca101d8771bd80f2bc4aaf1a8f19" />
+          <option name="title" value="新对话 2025年7月02日 15:35:43" />
+          <option name="updateTime" value="1751441743239" />
+        </Conversation>
+        <Conversation>
+          <option name="createTime" value="1751441398488" />
+          <option name="id" value="0197ca0adad875168de40d792dcb7b4c" />
+          <option name="title" value="新对话 2025年7月02日 15:29:58" />
+          <option name="updateTime" value="1751441398488" />
+        </Conversation>
         <Conversation>
           <option name="createTime" value="1750474299387" />
           <option name="id" value="0197906617fb7194a0407baae2b1e2eb" />

test_ori.py

@@ -14,6 +14,7 @@ from tools.dataset import get_transform
 from tools.image_joint import merge_imgs
 from configs import trainer_tools
 import yaml
+from datetime import datetime
 
 with open('configs/test.yml', 'r') as f:
     conf = yaml.load(f, Loader=yaml.FullLoader)

train_compare.py

@@ -5,6 +5,10 @@ import torch
 import torch.nn as nn
 import torch.optim as optim
 from tqdm import tqdm
+from torch.nn.parallel import DistributedDataParallel as DDP
+import torch.distributed as dist
+import torch.multiprocessing as mp
+import torch.utils.data.distributed
 
 from model.loss import FocalLoss
 from tools.dataset import load_data
@@ -12,136 +16,294 @@ import matplotlib.pyplot as plt
 from configs import trainer_tools
 import yaml
 from datetime import datetime
-with open('configs/scatter.yml', 'r') as f:
-    conf = yaml.load(f, Loader=yaml.FullLoader)
 
-# Data Setup
-train_dataloader, class_num = load_data(training=True, cfg=conf)
-val_dataloader, _ = load_data(training=False, cfg=conf)
+
+def load_config(config_path='configs/scatter.yml'):
+    """加载配置文件."""
+    with open(config_path, 'r') as f:
+        return yaml.load(f, Loader=yaml.FullLoader)
+
+
+# 加载配置
+conf = load_config()
+
+
+# 数据加载封装
+def load_datasets():
+    """加载训练和验证数据集，并为分布式训练创建 DistributedSampler."""
+    train_dataset, class_num = load_data(training=True, cfg=conf)
+    val_dataset, _ = load_data(training=False, cfg=conf)
+
+    if conf['base']['distributed']:
+        train_sampler = torch.utils.data.distributed.DistributedSampler(train_dataset)
+        val_sampler = torch.utils.data.distributed.DistributedSampler(val_dataset)
+    else:
+        train_sampler = None
+        val_sampler = None
+
+    train_dataloader = torch.utils.data.DataLoader(
+        train_dataset,
+        batch_size=conf['data']['batch_size'],
+        shuffle=(train_sampler is None),
+        num_workers=conf['data']['num_workers'],
+        pin_memory=conf['data']['pin_memory'],
+        sampler=train_sampler
+    )
+
+    val_dataloader = torch.utils.data.DataLoader(
+        val_dataset,
+        batch_size=conf['data']['batch_size'],
+        shuffle=False,
+        num_workers=conf['data']['num_workers'],
+        pin_memory=conf['data']['pin_memory'],
+        sampler=val_sampler
+    )
+
+    return train_dataloader, val_dataloader, class_num
+
+
+# 加载数据集
+train_dataloader, val_dataloader, class_num = load_datasets()
 
 tr_tools = trainer_tools(conf)
 backbone_mapping = tr_tools.get_backbone()
 metric_mapping = tr_tools.get_metric(class_num)
 
+
+# 设备管理封装
+def get_device(device_config=None):
+    """根据配置返回设备（CPU/GPU）。在分布式环境下初始化进程组."""
+    if device_config is None:
+        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+    else:
+        device = torch.device(device_config)
+
+    if conf['base']['distributed']:
+        dist.init_process_group(backend='nccl')
+
+    return device
+
+
+# 获取设备
+device = get_device(conf['base']['device'])
+
+# 模型初始化
 if conf['models']['backbone'] in backbone_mapping:
-    model = backbone_mapping[conf['models']['backbone']]().to(conf['base']['device'])
+    model = backbone_mapping[conf['models']['backbone']]().to(device)
 else:
     raise ValueError('不支持该模型: {}'.format({conf['models']['backbone']}))
 
 if conf['training']['metric'] in metric_mapping:
-    metric = metric_mapping[conf['training']['metric']]()
+    metric = metric_mapping[conf['training']['metric']]().to(device)
 else:
     raise ValueError('不支持的metric类型: {}'.format(conf['training']['metric']))
 
+rank = 0
 if torch.cuda.device_count() > 1 and conf['base']['distributed']:
     print("Let's use", torch.cuda.device_count(), "GPUs!")
-    model = nn.DataParallel(model)
-    metric = nn.DataParallel(metric)
+    dist.barrier()
+    model = DDP(model, device_ids=[rank])
+    metric = DDP(metric, device_ids=[rank])
 
 # Training Setup
-if conf['training']['loss'] == 'focal_loss':
-    criterion = FocalLoss(gamma=2)
-else:
-    criterion = nn.CrossEntropyLoss()
+def initialize_components():
+    # 封装模型、损失函数和优化器的初始化
+    if conf['training']['loss'] == 'focal_loss':
+        criterion = FocalLoss(gamma=2)
+    else:
+        criterion = nn.CrossEntropyLoss()
 
-optimizer_mapping = tr_tools.get_optimizer(model, metric)
-if conf['training']['optimizer'] in optimizer_mapping:
-    optimizer = optimizer_mapping[conf['training']['optimizer']]()
-    scheduler_mapping = tr_tools.get_scheduler(optimizer)
-    scheduler = scheduler_mapping[conf['training']['scheduler']]()
-    print('使用{}优化器 使用{}调度器'.format(conf['training']['optimizer'],
-                                             conf['training']['scheduler']))
+    optimizer_mapping = tr_tools.get_optimizer(model, metric)
+    if conf['training']['optimizer'] in optimizer_mapping:
+        optimizer = optimizer_mapping[conf['training']['optimizer']]()
+        scheduler_mapping = tr_tools.get_scheduler(optimizer)
+        scheduler = scheduler_mapping[conf['training']['scheduler']]()
+        print('使用{}优化器 使用{}调度器'.format(conf['training']['optimizer'],
+                                                 conf['training']['scheduler']))
+        return criterion, optimizer, scheduler
+    else:
+        raise ValueError('不支持的优化器类型: {}'.format(conf['training']['optimizer']))
 
-else:
-    raise ValueError('不支持的优化器类型: {}'.format(conf['training']['optimizer']))
+
+# 初始化组件
+criterion, optimizer, scheduler = initialize_components()
 
 # Checkpoints Setup
 checkpoints = conf['training']['checkpoints']
 os.makedirs(checkpoints, exist_ok=True)
 
+
+def train_epoch(model, dataloader, optimizer, criterion, device):
+    model.train()
+    train_loss = 0
+    for data, labels in tqdm(dataloader, desc="Training", ascii=True, total=len(dataloader)):
+        data, labels = data.to(device), labels.to(device)
+        embeddings = model(data).to(device)
+        if not conf['training']['metric'] == 'softmax':
+            thetas = metric(embeddings, labels)
+        else:
+            thetas = metric(embeddings)
+        loss = criterion(thetas, labels)
+        optimizer.zero_grad()
+        loss.backward()
+        optimizer.step()
+        train_loss += loss.item()
+    return train_loss / len(dataloader)
+
+
+def validate_epoch(model, dataloader, criterion, device):
+    model.eval()
+    val_loss = 0
+    with torch.no_grad():
+        for data, labels in tqdm(dataloader, desc="Validation", ascii=True, total=len(dataloader)):
+            data, labels = data.to(device), labels.to(device)
+            embeddings = model(data).to(device)
+            if not conf['training']['metric'] == 'softmax':
+                thetas = metric(embeddings, labels)
+            else:
+                thetas = metric(embeddings)
+            loss = criterion(thetas, labels)
+            val_loss += loss.item()
+    return val_loss / len(dataloader)
+
+
+def save_model(model, path, distributed):
+    if distributed and torch.cuda.device_count() > 1:
+        if dist.get_rank() == 0:
+            torch.save(model.module.state_dict(), path)
+    else:
+        torch.save(model.state_dict(), path)
+
+
+def write_log(log_info, log_dir):
+    with open(log_dir, 'a') as f:
+        f.write(log_info + '\n')
+
+
+def plot_losses(epochs, train_losses, val_losses, save_path):
+    plt.plot(epochs, train_losses, color='blue', label='Train Loss')
+    plt.plot(epochs, val_losses, color='red', label='Validation Loss')
+    plt.xlabel('Epochs')
+    plt.ylabel('Loss')
+    plt.legend()
+    plt.savefig(save_path)
+    plt.close()
+
+
+# 模型恢复封装
+def restore_model(model, device):
+    if conf['training']['restore']:
+        print('load pretrain model: {}'.format(conf['training']['restore_model']))
+        model.load_state_dict(torch.load(conf['training']['restore_model'], map_location=device))
+    return model
+
+
+# 日志和学习率记录封装
+def log_and_print(e, train_loss_avg, val_loss_avg, current_lr, log_dir):
+    if conf['base']['distributed'] and dist.get_rank() != 0:
+        return
+
+    log_info = ("[{:%Y-%m-%d %H:%M:%S}] Epoch {}/{},  train_loss: {}, val_loss: {} lr:{}"
+                .format(datetime.now(),
+                        e,
+                        conf['training']['epochs'],
+                        train_loss_avg,
+                        val_loss_avg,
+                        current_lr))
+    print(log_info)
+    write_log(log_info, log_dir)
+    print("第%d个epoch的学习率：%f" % (e, current_lr))
+
+
+# 模型评估与保存封装
+def evaluate_and_save(val_loss_avg, best_loss, model, checkpoints, distributed):
+    if val_loss_avg < best_loss:
+        best_path = osp.join(checkpoints, 'best.pth')
+        save_model(model, best_path, distributed)
+        best_loss = val_loss_avg
+    return best_loss
+
+
+def run_training(rank, world_size, conf):
+    """在指定 rank 上运行训练 loop。
+
+    Args:
+        rank: 当前进程的索引。
+        world_size: 进程总数。
+        conf: 配置字典。
+    """
+    dist.init_process_group(backend='nccl', rank=rank, world_size=world_size)
+    device = torch.device('cuda', rank)
+
+    # 数据加载器和模型等需要重新初始化以确保每个进程独立工作
+    train_dataloader, val_dataloader, class_num = load_datasets()
+    tr_tools = trainer_tools(conf)
+    backbone_mapping = tr_tools.get_backbone()
+    metric_mapping = tr_tools.get_metric(class_num)
+
+    # 模型初始化
+    if conf['models']['backbone'] in backbone_mapping:
+        model = backbone_mapping[conf['models']['backbone']]().to(device)
+    else:
+        raise ValueError('不支持该模型: {}'.format({conf['models']['backbone']}))
+
+    if conf['training']['metric'] in metric_mapping:
+        metric = metric_mapping[conf['training']['metric']]().to(device)
+    else:
+        raise ValueError('不支持的metric类型: {}'.format(conf['training']['metric']))
+
+    model = DDP(model, device_ids=[rank])
+    metric = DDP(metric, device_ids=[rank])
+
+    # 初始化组件
+    criterion, optimizer, scheduler = initialize_components()
+
+    # 恢复模型（如果需要）
+    model = restore_model(model, device)
+
+    train_losses = []
+    val_losses = []
+    epochs = []
+    temp_loss = 1000
+    log_dir = osp.join(conf['logging']['logging_dir'])
+
+    for e in range(conf['training']['epochs']):
+        train_loss_avg = train_epoch(model, train_dataloader, optimizer, criterion, device)
+        train_losses.append(train_loss_avg)
+
+        val_loss_avg = validate_epoch(model, val_dataloader, criterion, device)
+        val_losses.append(val_loss_avg)
+
+        temp_loss = evaluate_and_save(val_loss_avg, temp_loss, model, conf['training']['checkpoints'], True)
+
+        scheduler.step()
+        current_lr = optimizer.param_groups[0]['lr']
+        log_and_print(e, train_loss_avg, val_loss_avg, current_lr, log_dir)
+        epochs.append(e)
+
+    last_path = osp.join(conf['training']['checkpoints'], 'last.pth')
+    save_model(model, last_path, True)
+    plot_losses(epochs, train_losses, val_losses, 'loss/mobilenetv3Large_2250_0316.png')
+
+    dist.destroy_process_group()
+
+
+def main():
+    world_size = torch.cuda.device_count()
+    mp.spawn(
+        run_training,
+        args=(world_size, conf),
+        nprocs=world_size,
+        join=True,
+    )
+
+
 if __name__ == '__main__':
     print('backbone>{}  '.format(conf['models']['backbone']),
           'metric>{}  '.format(conf['training']['metric']),
           'checkpoints>{}  '.format(conf['training']['checkpoints']),
           )
-    train_losses = []
-    val_losses = []
-    epochs = []
-    temp_loss = 100
-    if conf['training']['restore']:
-        print('load pretrain model: {}'.format(conf['training']['restore_model']))
-        model.load_state_dict(torch.load(conf['training']['restore_model'],
-                                         map_location=conf['base']['device']))
-
-    for e in range(conf['training']['epochs']):
-        train_loss = 0
-        model.train()
-
-        for train_data, train_labels in tqdm(train_dataloader,
-                                             desc="Epoch {}/{}"
-                                                     .format(e, conf['training']['epochs']),
-                                             ascii=True,
-                                             total=len(train_dataloader)):
-            train_data = train_data.to(conf['base']['device'])
-            train_labels = train_labels.to(conf['base']['device'])
-
-            train_embeddings = model(train_data).to(conf['base']['device'])  # [256,512]
-            # pdb.set_trace()
-
-            if not conf['training']['metric'] == 'softmax':
-                thetas = metric(train_embeddings, train_labels)  # [256,357]
-            else:
-                thetas = metric(train_embeddings)
-            tloss = criterion(thetas, train_labels)
-            optimizer.zero_grad()
-            tloss.backward()
-            optimizer.step()
-            train_loss += tloss.item()
-        train_lossAvg = train_loss / len(train_dataloader)
-        train_losses.append(train_lossAvg)
-        epochs.append(e)
-        val_loss = 0
-        model.eval()
-        with torch.no_grad():
-            for val_data, val_labels in tqdm(val_dataloader, desc="val",
-                                             ascii=True, total=len(val_dataloader)):
-                val_data = val_data.to(conf['base']['device'])
-                val_labels = val_labels.to(conf['base']['device'])
-                val_embeddings = model(val_data).to(conf['base']['device'])
-                if not conf['training']['metric'] == 'softmax':
-                    thetas = metric(val_embeddings, val_labels)
-                else:
-                    thetas = metric(val_embeddings)
-                vloss = criterion(thetas, val_labels)
-                val_loss += vloss.item()
-            val_lossAvg = val_loss / len(val_dataloader)
-            val_losses.append(val_lossAvg)
-            if val_lossAvg < temp_loss:
-                if torch.cuda.device_count() > 1:
-                    torch.save(model.state_dict(), osp.join(checkpoints, 'best.pth'))
-                else:
-                    torch.save(model.state_dict(), osp.join(checkpoints, 'best.pth'))
-                temp_loss = val_lossAvg
-
-        scheduler.step()
-        current_lr = optimizer.param_groups[0]['lr']
-        log_info = ("[{:%Y-%m-%d %H:%M:%S}] Epoch {}/{},  train_loss: {}, val_loss: {} lr:{}"
-                    .format(datetime.now(),
-                            e,
-                            conf['training']['epochs'],
-                            train_lossAvg,
-                            val_lossAvg,
-                            current_lr))
-        print(log_info)
-        # 写入日志文件
-        with open(osp.join(conf['logging']['logging_dir']), 'a') as f:
-            f.write(log_info + '\n')
-        print("第%d个epoch的学习率：%f" % (e, current_lr))
-    if torch.cuda.device_count() > 1 and conf['base']['distributed']:
-        torch.save(model.module.state_dict(), osp.join(checkpoints, 'last.pth'))
+    if conf['base']['distributed']:
+        main()
     else:
-        torch.save(model.state_dict(), osp.join(checkpoints, 'last.pth'))
-    plt.plot(epochs, train_losses, color='blue')
-    plt.plot(epochs, val_losses, color='red')
-    # plt.savefig('lossMobilenetv3.png')
-    plt.savefig('loss/mobilenetv3Large_2250_0316.png')
+        run_training(rank=0, world_size=1, conf=conf)

train_compare.py.bak

@@ -0,0 +1,147 @@
+import os
+import os.path as osp
+
+import torch
+import torch.nn as nn
+import torch.optim as optim
+from tqdm import tqdm
+
+from model.loss import FocalLoss
+from tools.dataset import load_data
+import matplotlib.pyplot as plt
+from configs import trainer_tools
+import yaml
+from datetime import datetime
+with open('configs/scatter.yml', 'r') as f:
+    conf = yaml.load(f, Loader=yaml.FullLoader)
+
+# Data Setup
+train_dataloader, class_num = load_data(training=True, cfg=conf)
+val_dataloader, _ = load_data(training=False, cfg=conf)
+
+tr_tools = trainer_tools(conf)
+backbone_mapping = tr_tools.get_backbone()
+metric_mapping = tr_tools.get_metric(class_num)
+
+if conf['models']['backbone'] in backbone_mapping:
+    model = backbone_mapping[conf['models']['backbone']]().to(conf['base']['device'])
+else:
+    raise ValueError('不支持该模型: {}'.format({conf['models']['backbone']}))
+
+if conf['training']['metric'] in metric_mapping:
+    metric = metric_mapping[conf['training']['metric']]()
+else:
+    raise ValueError('不支持的metric类型: {}'.format(conf['training']['metric']))
+
+if torch.cuda.device_count() > 1 and conf['base']['distributed']:
+    print("Let's use", torch.cuda.device_count(), "GPUs!")
+    model = nn.DataParallel(model)
+    metric = nn.DataParallel(metric)
+
+# Training Setup
+if conf['training']['loss'] == 'focal_loss':
+    criterion = FocalLoss(gamma=2)
+else:
+    criterion = nn.CrossEntropyLoss()
+
+optimizer_mapping = tr_tools.get_optimizer(model, metric)
+if conf['training']['optimizer'] in optimizer_mapping:
+    optimizer = optimizer_mapping[conf['training']['optimizer']]()
+    scheduler_mapping = tr_tools.get_scheduler(optimizer)
+    scheduler = scheduler_mapping[conf['training']['scheduler']]()
+    print('使用{}优化器 使用{}调度器'.format(conf['training']['optimizer'],
+                                             conf['training']['scheduler']))
+
+else:
+    raise ValueError('不支持的优化器类型: {}'.format(conf['training']['optimizer']))
+
+# Checkpoints Setup
+checkpoints = conf['training']['checkpoints']
+os.makedirs(checkpoints, exist_ok=True)
+
+if __name__ == '__main__':
+    print('backbone>{}  '.format(conf['models']['backbone']),
+          'metric>{}  '.format(conf['training']['metric']),
+          'checkpoints>{}  '.format(conf['training']['checkpoints']),
+          )
+    train_losses = []
+    val_losses = []
+    epochs = []
+    temp_loss = 100
+    if conf['training']['restore']:
+        print('load pretrain model: {}'.format(conf['training']['restore_model']))
+        model.load_state_dict(torch.load(conf['training']['restore_model'],
+                                         map_location=conf['base']['device']))
+
+    for e in range(conf['training']['epochs']):
+        train_loss = 0
+        model.train()
+
+        for train_data, train_labels in tqdm(train_dataloader,
+                                             desc="Epoch {}/{}"
+                                                     .format(e, conf['training']['epochs']),
+                                             ascii=True,
+                                             total=len(train_dataloader)):
+            train_data = train_data.to(conf['base']['device'])
+            train_labels = train_labels.to(conf['base']['device'])
+
+            train_embeddings = model(train_data).to(conf['base']['device'])  # [256,512]
+            # pdb.set_trace()
+
+            if not conf['training']['metric'] == 'softmax':
+                thetas = metric(train_embeddings, train_labels)  # [256,357]
+            else:
+                thetas = metric(train_embeddings)
+            tloss = criterion(thetas, train_labels)
+            optimizer.zero_grad()
+            tloss.backward()
+            optimizer.step()
+            train_loss += tloss.item()
+        train_lossAvg = train_loss / len(train_dataloader)
+        train_losses.append(train_lossAvg)
+        epochs.append(e)
+        val_loss = 0
+        model.eval()
+        with torch.no_grad():
+            for val_data, val_labels in tqdm(val_dataloader, desc="val",
+                                             ascii=True, total=len(val_dataloader)):
+                val_data = val_data.to(conf['base']['device'])
+                val_labels = val_labels.to(conf['base']['device'])
+                val_embeddings = model(val_data).to(conf['base']['device'])
+                if not conf['training']['metric'] == 'softmax':
+                    thetas = metric(val_embeddings, val_labels)
+                else:
+                    thetas = metric(val_embeddings)
+                vloss = criterion(thetas, val_labels)
+                val_loss += vloss.item()
+            val_lossAvg = val_loss / len(val_dataloader)
+            val_losses.append(val_lossAvg)
+            if val_lossAvg < temp_loss:
+                if torch.cuda.device_count() > 1:
+                    torch.save(model.state_dict(), osp.join(checkpoints, 'best.pth'))
+                else:
+                    torch.save(model.state_dict(), osp.join(checkpoints, 'best.pth'))
+                temp_loss = val_lossAvg
+
+        scheduler.step()
+        current_lr = optimizer.param_groups[0]['lr']
+        log_info = ("[{:%Y-%m-%d %H:%M:%S}] Epoch {}/{},  train_loss: {}, val_loss: {} lr:{}"
+                    .format(datetime.now(),
+                            e,
+                            conf['training']['epochs'],
+                            train_lossAvg,
+                            val_lossAvg,
+                            current_lr))
+        print(log_info)
+        # 写入日志文件
+        with open(osp.join(conf['logging']['logging_dir']), 'a') as f:
+            f.write(log_info + '\n')
+        print("第%d个epoch的学习率：%f" % (e, current_lr))
+    if torch.cuda.device_count() > 1 and conf['base']['distributed']:
+        torch.save(model.module.state_dict(), osp.join(checkpoints, 'last.pth'))
+    else:
+        torch.save(model.state_dict(), osp.join(checkpoints, 'last.pth'))
+    plt.plot(epochs, train_losses, color='blue')
+    plt.plot(epochs, val_losses, color='red')
+    # plt.savefig('lossMobilenetv3.png')
+    plt.savefig('loss/mobilenetv3Large_2250_0316.png')