Raise exceptions instead of asserts in...

Raise exceptions instead of asserts in src/transformers/models/bart/modeling_flax_[bart, marian, mbart, pegasus].py (#13939)

* Raise exceptions instead of asserts

* fix: fixed failing quality check with copies

* fix: fixed max line length

* rerun github ci, failed to install dependencies

src/transformers/models/bart/modeling_flax_bart.py

@@ -237,9 +237,11 @@ class FlaxBartAttention(nn.Module):
 
     def setup(self) -> None:
         self.head_dim = self.embed_dim // self.num_heads
-        assert (
-            self.head_dim * self.num_heads == self.embed_dim
-        ), f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`: {self.num_heads})."
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim}"
+                f" and `num_heads`: {self.num_heads})."
+            )
 
         dense = partial(
             nn.Dense,

src/transformers/models/marian/modeling_flax_marian.py

@@ -241,9 +241,11 @@ class FlaxMarianAttention(nn.Module):
 
     def setup(self) -> None:
         self.head_dim = self.embed_dim // self.num_heads
-        assert (
-            self.head_dim * self.num_heads == self.embed_dim
-        ), f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`: {self.num_heads})."
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim}"
+                f" and `num_heads`: {self.num_heads})."
+            )
 
         dense = partial(
             nn.Dense,

src/transformers/models/mbart/modeling_flax_mbart.py

@@ -248,9 +248,11 @@ class FlaxMBartAttention(nn.Module):
 
     def setup(self) -> None:
         self.head_dim = self.embed_dim // self.num_heads
-        assert (
-            self.head_dim * self.num_heads == self.embed_dim
-        ), f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`: {self.num_heads})."
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim}"
+                f" and `num_heads`: {self.num_heads})."
+            )
 
         dense = partial(
             nn.Dense,

src/transformers/models/pegasus/modeling_flax_pegasus.py

@@ -241,9 +241,11 @@ class FlaxPegasusAttention(nn.Module):
 
     def setup(self) -> None:
         self.head_dim = self.embed_dim // self.num_heads
-        assert (
-            self.head_dim * self.num_heads == self.embed_dim
-        ), f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`: {self.num_heads})."
+        if self.head_dim * self.num_heads != self.embed_dim:
+            raise ValueError(
+                f"embed_dim must be divisible by num_heads (got `embed_dim`: {self.embed_dim}"
+                f" and `num_heads`: {self.num_heads})."
+            )
 
         dense = partial(
             nn.Dense,