Validation Set

Early Stopping

अर्ली स्टपिङ

Halting training once validation performance stops improving to avoid overfitting.

Dropout

ड्रपआउट

A regularization method that randomly disables neurons during training to reduce overfitting.

Checkpoint

चेकपोइन्ट

A saved snapshot of a model's state during training so it can be resumed or reused.

Gradient Clipping

ग्र्याडियन्ट क्लिपिङ

Capping gradient size during training to prevent unstable, exploding updates.

Hyperparameter

हाइपरप्यारामिटर

A configuration value set before training that controls how a model learns, like learning rate.

Overfitting

ओभरफिटिङ / अति-अनुकूलन

When a model learns the training data too closely, including noise, and fails on new data.

Parameter

प्यारामिटर

A value inside a model (like a weight) learned from data during training.

Pre-training

पूर्व-तालिम

The initial large-scale training of a model on broad data before any task-specific tuning.