Marian MT model (#1210)

* Skeleton files for the marian MT model.

* Marian initialization.

* Implement the attention forward method.

* Forward pass for the encoder side.

* Expose the encoder and decoder.

* Start plugging the decoder.

* Forward pass for the decoder layer.

* Set up the marian example.

* Add some missing backtraces.

* Bugfix.

candle-transformers/src/models/with_tracing.rs

@@ -14,6 +14,13 @@ impl Embedding {
         Ok(Self { inner, span })
     }
 
+    pub fn from_weights(weights: Tensor) -> Result<Self> {
+        let (_in_size, out_size) = weights.dims2()?;
+        let inner = candle_nn::Embedding::new(weights, out_size);
+        let span = tracing::span!(tracing::Level::TRACE, "embedding");
+        Ok(Self { inner, span })
+    }
+
     pub fn embeddings(&self) -> &Tensor {
         self.inner.embeddings()
     }