How do you track validation loss, while finetuning the SmolVLM using the finetuning notebook given.

[Aritra02091998]

I assigned a validation dataset in the training arguments and set the following in the TrainingArguments:

eval_strategy="steps",        # Enables evaluation
eval_steps=2,                 # Frequency of evaluation
load_best_model_at_end=True,  # Load the best model at the end
metric_for_best_model="eval_loss",  # Monitor this metric

I didn't make nay changes except these to the finetuning code. But the eval_loss is not being calculated. And showing me this error:

KeyError: "The metric_for_best_model training argument is set to 'eval_loss', which is not found in the evaluation metrics. The available evaluation metrics are: []. Please ensure that the compute_metrics function returns a dictionary that includes 'eval_loss' or consider changing the metric_for_best_model via the TrainingArguments."

How can I fix this ? Any leads will be helpful.