Fig. 2From: Lerna: transformer architectures for configuring error correction tools for short- and long-read genome sequencingVariation of perplexity and % of aligned reads with k on simulated E. coli PacBio reads, corrected by LoRDEC, mapped by Minimap2. The perplexity is computed by a Transformer word-level LM with \(|w|=4\). A strong negative correlation between the two metrics is clearly observedBack to article page