RCN implementation on a small train and test set

[shrinuKushagra]

This PR contains the first implementation of the RCN example using the PGMax package.
The file to run is examples/rcn/inference_pgmax_small.py
This code contains implementation, visualization on a small set. Trained with 20 examples and tested on 20 examples.

The inference has been separated from model creation code.
Saved models are added to /storage/users/skushagra/pgmax_rcn_artifacts/ .

Implementation on the full dataset will be implemented in a later PR.

[codecov-commenter]

Codecov Report

Merging #93 (031e1ed) into master (2f389ab) will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##            master       #93   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            8         8           
  Lines          631       631           
=========================================
  Hits           631       631           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2f389ab...031e1ed. Read the comment docs.

[StannisZhou]

Can we get the (standalone) notebook in a separate PR? The notebook should be self-contained (without reference to other files). We can keep this PR for later evaluation on the full test set.

[StannisZhou]

We can easily construct these phis on demand so probably better to construct them instead of loading them from disk.

[shrinuKushagra]

Computing phi involves the following code

rows = []
cols = []
index = 0
for i in range(1, M):
    r1, c1 = index_to_rc(i, hps, vps)

    r2_min = max(r1 - r, -hps)
    r2_max = min(r1 + r, hps)
    c2_min = max(c1 - r, -vps)
    c2_max = min(c1 + r, vps)

    for r2 in range(r2_min, r2_max + 1):
        for c2 in range(c2_min, c2_max + 1):
            j = rc_to_index(r2, c2, hps, vps)
            rows.append(i)
            cols.append(j)
            index += 1

which is not too complicated. For simplicity of the current notebook; I didnt include it in the example. We can include it if you think it helps the presentation.

[NishanthJKumar]

This PR seems stale and redundant given #96 has been merged. Should we close?

[StannisZhou]

This PR seems stale and redundant given #96 has been merged. Should we close?
Good call. Closed!