SFTTrainer not reporting training loss

Beginners
1

I am using SFTTrainer with Lora to fine tune gemma model. I am getting validation loss but training loss is reported as “No log”

Here is my configuration:

args = TrainingArguments(
        per_device_train_batch_size=6,
        per_device_eval_batch_size=6,
        gradient_accumulation_steps=4,
        save_steps=200,
        save_total_limit=30,
        save_strategy="steps",
        evaluation_strategy='steps',
        eval_steps=200,
        logging_steps=200,
    
        warmup_steps=2,
        num_train_epochs =2,
        # max_steps=2,
        learning_rate=2e-4,
        lr_scheduler_type = "cosine",
        weight_decay = 0.001,
        max_grad_norm=1.0,
        fp16 = False,
        bf16 = True,
        logging_strategy = "epoch",
        output_dir=output_dir,
        optim="paged_adamw_8bit",
        seed=42,
        gradient_checkpointing = True,
        gradient_checkpointing_kwargs={'use_reentrant':False},
        #accelerator_config = {'split_batches' : True},
        report_to = None
    )

trainer = SFTTrainer(
    model=model,
    train_dataset=dataset['train'],
    eval_dataset=dataset["test"] , 
    args=args ,
    peft_config=lora_config,
    dataset_text_field=dataset_text_field,
    max_seq_length = 512,
    tokenizer = tokenizer,
    callbacks=[SavePeftModelCallback],# This is crashing the kernel
    # data_collator=collator,
)

image