Add adapted model

README.md

@@ -8,6 +8,10 @@ license: apache-2.0
 This model contains the converted PyTorch checkpoint of the original Tensorflow model available in the [TaPas repository](https://github.com/google-research/tapas/blob/master/DENSE_TABLE_RETRIEVER.md#reader-models).
 It is described in Herzig et al.'s (2021) [paper](https://aclanthology.org/2021.naacl-main.43/) _Open Domain Question Answering over Tables via Dense Retrieval_.
 
+This model has 2 versions which can be used differing only in the table scoring head.
+The default one has an adapted table scoring head in order to be able to generate probabilities out of the logits.
+The other (non-default) version corredponds to the original checkpoint from the TaPas repository and can be accessed setting `revision="original"`.
+
 
 # Usage
 ## In Haystack

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4267f2a38e29c2e2e9fadef32ab4a70a81a07ffeaf9b554dfa1d2ce2615ccee6
-size 1347084063
+oid sha256:5fbde7b8e6044cec61f9f8e4d4c2d731c3b98a57a8c9ab3b02fbf82c7556ce0e
+size 1347080931