trocr viarant for Predictor runtime: no model size for input_shape…

Because transformers v4 and v5 API for image preprocessor differs,
and the model-internal image input sizes are actually irrelevant,
because the preprocessor will resize them anyway, and there is no
batch dimension (because the input images will have different shapes),
do not advertise this information in `.input_shape`.

src/eynollah/model_zoo/model_zoo.py

@@ -379,9 +379,9 @@ class EynollahModelZoo:
             return text, conf
         model.predict_on_batch = predict_torch
         # not actually needed (image processor does resize itself)
+        # no batch dimension (images passed as list w/ varying shapes)
         model.input_shape = (None,
-                             proc.image_processor.size.height,
-                             proc.image_processor.size.width,
+                             None,
                              len(proc.image_processor.image_mean))
         return model