one

done

config.json

@@ -76,3 +76,10 @@
         "type": "learned_abs_2d",
         "max_pos_embeddings": 50
     },
+    "image_feature_source": ["spatial_avg_pool", "temporal_avg_pool"]
+  },
+  "vocab_size": 51289,
+  "torch_dtype": "float16",
+  "transformers_version": "4.41.0.dev0",
+  "is_encoder_decoder": true
+}