summaryrefslogtreecommitdiff
path: root/research/lesci/extract.go
diff options
context:
space:
mode:
Diffstat (limited to 'research/lesci/extract.go')
-rw-r--r--research/lesci/extract.go12
1 files changed, 1 insertions, 11 deletions
diff --git a/research/lesci/extract.go b/research/lesci/extract.go
index c7c85e8..659cdc3 100644
--- a/research/lesci/extract.go
+++ b/research/lesci/extract.go
@@ -52,17 +52,7 @@ func Rules(tokenizer llm.Tokenizer, merges [][2]string) (tensor.Dense[int64], []
rules := tensor.NewDense[int64]([]int{len(merges), 3}, nil)
- atoi := make(map[string]int)
-
- vocab := bpe.Vocab(tokenizer.(*bpe.Tokenizer))
-
- for i, token := range vocab {
- if _, ok := atoi[token]; ok {
- continue
- }
-
- atoi[token] = i
- }
+ atoi := bpe.Atoi(tokenizer.(*bpe.Tokenizer))
for i, merge := range merges {
if !valid[i] {