# tinker.types.AdamParams

## *class* [**tinker.types.AdamParams**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L12)(*[StrictBase](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/_models.py#L76)*)

**Fields:**

- [**learning_rate**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L13) (*float*) – Learning rate for the optimizer
- [**beta1**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L16) (*float*) – Coefficient used for computing running averages of gradient
- [**beta2**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L19) (*float*) – Coefficient used for computing running averages of gradient square
- [**eps**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L22) (*float*) – Term added to the denominator to improve numerical stability
- [**weight_decay**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L25) (*float*) – Weight decay for the optimizer. Uses decoupled weight decay.
- [**grad_clip_norm**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/optim_step_request.py#L28) (*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.

## Referenced by

- [TrainingClient.optim_step](https://tinker-docs.thinkingmachines.ai/tinker/api-reference/trainingclient/#optim_step)