Fix gemma3_text tokenizer in mapping (#36793)

Author: LysandreJik

src/transformers/models/auto/tokenization_auto.py

@@ -222,6 +222,13 @@
                     "GemmaTokenizerFast" if is_tokenizers_available() else None,
                 ),
             ),
+            (
+                "gemma3_text",
+                (
+                    "GemmaTokenizer" if is_sentencepiece_available() else None,
+                    "GemmaTokenizerFast" if is_tokenizers_available() else None,
+                ),
+            ),
             ("git", ("BertTokenizer", "BertTokenizerFast" if is_tokenizers_available() else None)),
             ("glm", (None, "PreTrainedTokenizerFast" if is_tokenizers_available() else None)),
             ("gpt-sw3", ("GPTSw3Tokenizer" if is_sentencepiece_available() else None, None)),