Gradient Clipping

Checkpoint

चेकपोइन्ट

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

Dropout

ड्रपआउट

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

Parameter

प्यारामिटर

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

Stochastic Gradient Descent (SGD)

स्टोकास्टिक ग्र्याडियन्ट डिसेन्ट

Gradient descent that updates weights using small random batches for speed.

Validation Set

प्रमाणीकरण सेट

Data used during training to tune settings and avoid overfitting before final testing.

Vanishing Gradient

भ्यानिसिङ ग्र्याडियन्ट

A problem where learning signals shrink in deep networks, slowing or stalling training.

Bagging

ब्यागिङ

Training many models on random data subsets and averaging them to reduce variance.

Batch

ब्याच

A small subset of training data processed together in one step of training.