Add training for the llama2.c example (#296)

* Rework the commands and run inference by default.

* Add the training module and load the training dataset.

* Random dataset iterator.

* Proper valid-loss computation.

* Compute the evaluation loss.

* Add more substance to the training loop.

candle-core/src/lib.rs

@@ -44,7 +44,7 @@ mod device;
 pub mod display;
 mod dtype;
 mod dummy_cuda_backend;
-mod error;
+pub mod error;
 mod indexer;
 pub mod layout;
 #[cfg(feature = "mkl")]