AdamParams - Tinker Documentation
tinker.types.AdamParams
class tinker.types.AdamParams( StrictBase)
Fields:
- learning_rate ( float) – Learning rate for the optimizer
- beta1 ( float) – Coefficient used for computing running averages of gradient
- beta2 ( float) – Coefficient used for computing running averages of gradient square
- eps ( float) – Term added to the denominator to improve numerical stability
- weight_decay ( float) – Weight decay for the optimizer. Uses decoupled weight decay.
- grad_clip_norm ( float) – Maximum global gradient norm. If the global gradient norm is greater than this value, it will be clipped to this value. 0.0 means no clipping.