Fix MS MARCO v1 doc segmented uniCOIL encoding issues

[lintool]

Background, ref #1850

On MS MARCO v2, in addition to these previous corpora:

889db095113cc4fe152382ccff73304a  msmarco_v2_doc_segmented_unicoil_0shot.tar
28261587d6afde56efd8df4f950e7fb4  msmarco_v2_doc_segmented_unicoil_noexp_0shot.tar

@MXueguang also created these:

c5639748c2cbad0152e10b0ebde3b804  msmarco_v2_doc_segmented_unicoil_0shot_v2.tar
97ba262c497164de1054f357caea0c63  msmarco_v2_doc_segmented_unicoil_noexp_0shot_v2.tar

The first set is encoded only with segment; the second prepended doc title.

We wanted do the same with V1. These are the existing corpora:

d7536219c03cc24c60654bea18204e93  msmarco-doc-segmented-unicoil-noexp.tar
6a00e2c0c375cb1e52c83ae5ac377ebb  msmarco-doc-segmented-unicoil.tar

And these are the new ones @MXueguang prepared:

bc71c588c9a2ed175cf2fbe52bd0b86e  msmarco-doc-segmented-unicoil-noexp-v2.tar
3618700294e0edc59b2b3b0820e3acf9  msmarco-doc-segmented-unicoil-v2.tar

However, we later discovered:

1. msmarco-passage-unicoil-noexp.tar - appears to be just segment encoded
2. msmarco-passage-unicoil.tar - actually appears to be title/segment encoded
3. msmarco-passage-unicoil-noexp-v2.tar- title/segment encoded
4. msmarco-passage-unicoil-v2.tar - title/segment encoded

So, it appears that (2) and (4) are the same condition.

Experimental results check out - (3) is more effective than (1), but (4) ~ (2), differences due to noise.

Here's my plan - let's replace (1) with (3) - i.e., just overwrite, but retain (2); so we'll discard (4). Thus, we will not be creating new YAML files for regressions.

• msmarco-passage-unicoil: remains exactly the same
• msmarco-passage-unicoil-noexp: updated to title/segment encoding