Training loss remained almost unchanged during training

[Hi_Hieu]

Hi.
I am training a model.
This is my code to train the model.

    optimizer = optim.SGD(model.parameters(), lr=args.learning_rate)
    criterion = InfoNCELoss(temperature=0.1)
    scheduler = optim.lr_scheduler.StepLR(optimizer, 1.0, gamma=0.1)

    log_iter = 100
    
    for e in range(1, args.epoch + 1):
        model.train()
        t0 = time.time()
        total_train_loss = 0

        for i, batch in enumerate(train_dataloader):
 
            optimizer.zero_grad()
            sentence1_embedding, sentence2_embedding = model(batch["sentence1"], batch["sentence2"])
            train_batch_loss = criterion(sentence1_embedding, sentence2_embedding, batch["sentence1_label"], batch["sentence2_label"])

            total_train_loss += train_batch_loss.item()
            train_batch_loss.backward()
            optimizer.step()
            scheduler.step()

This is my training loss:

Training loss remained almost unchanged during training.
Please help me to solve this.

Thank!