[caffe2] Fix DBFileReader (#53498)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53498

This code depended on `Blobs()` being returned in sorted order:

https://www.internalfb.com/intern/diffusion/FBS/browsefile/master/fbcode/caffe2/caffe2/python/db_file_reader.py?commit=472774e7f507e124392491800d9654e01269cbaf&lines=89-91

But D26504408 (https://github.com/pytorch/pytorch/commit/69bb0e028596d481cda38a0f49bcb5965fab4293) changed the underlying storage to a hashmap, so now the blobs are returned in arbitrary order (Note that `Blobs()` returns also non-local blobs, and for those there was already no guarantee of ordering).

So we need to explicitly sort the result.

Test Plan:
```
$ buck test dper3/dper3/toolkit/tests:lime_test
$ buck test //dper3/dper3/toolkit/tests:model_insight_test
```
Pass after this diff.

Differential Revision: D26879502

fbshipit-source-id: d76113f8780544af1d97ec0a818fb21cc767f2bf
diff --git a/caffe2/python/db_file_reader.py b/caffe2/python/db_file_reader.py
index 265b192..49b1609 100644
--- a/caffe2/python/db_file_reader.py
+++ b/caffe2/python/db_file_reader.py
@@ -118,7 +118,7 @@
             )
         )
         col_names = [
-            blob_name[len(blob_prefix):] for blob_name in workspace.Blobs()
+            blob_name[len(blob_prefix):] for blob_name in sorted(workspace.Blobs())
             if blob_name.startswith(blob_prefix)
         ]
         schema = from_column_list(col_names)