Laurent Mazare ff876c2103 Llama more training (#297) ...

* Rework the var-builder to handle initializations.

* Add some helper functions for layer creation.

* Improve the layer initializations.

* Get initialized variables.

* Precompute the rot embeddings when training lamas.

2023-08-01 19:53:41 +01:00

..

layer_norm.rs

Simplify the parameters used by sum and sum_keepdim. (#165)

2023-07-14 08:22:08 +01:00

loss.rs

Add the cross-entropy loss. (#287)

2023-07-31 14:26:36 +01:00

ops.rs

Softmax numerical stability. (#267)

2023-07-28 13:13:01 +01:00

optim.rs

Llama more training (#297)

2023-08-01 19:53:41 +01:00