tballison commented on code in PR #2878:
URL: https://github.com/apache/tika/pull/2878#discussion_r3367368110
##########
tika-ml/tika-ml-junkdetect/src/main/java/org/apache/tika/ml/junkdetect/tools/TrainJunkModel.java:
##########
@@ -1037,12 +1037,8 @@ public static BigramTables buildBigramTablesFromCounts(
// Quantize unigram log-probs.
QuantizedFloats qUnigram = quantizeFloats(unigramLogP);
- // --- Build the open-addressing bigram table. ---
- int slots = nextPowerOfTwo((int) Math.max(2, Math.ceil(keptPairs /
loadFactor)));
- int[] keys = new int[slots];
- java.util.Arrays.fill(keys, BigramTables.EMPTY_KEY);
- // Compute log-probs first, quantize once, then write into the table
- // alongside its key.
+ // --- Build the sorted-occupied bigram table (binary-search lookup).
---
+ // Compute log-probs first, quantize once, then sort by key.
float[] keptLogP = new float[keptPairs];
Review Comment:
good catch
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
