Lower Mem AGKM + Variable Len Seqlet ID + Exemplar-Based Hit Scoring

[AvantiShri]

• AGKM embeddings now default (also avoids importing tensorflow unless user wants to use the previous gapped kmer embeddings)
• cap on the agkm embeddings size; reduces memory use. See https://github.com/kundajelab/tfmodisco_bio_experiments/blob/4126634304693688f2ecce1032ccee62b9eb3889/bpnet/trial1/Trying_Advanced_GappedKmerEmbedding.ipynb for exploration of different cap sizes.
• Implementation of variable-length seqlet identification
• Exemplar-based hit-scoring