From ce9c65cb1e13f7ae6eed4b419232a4da9dd63fdb Mon Sep 17 00:00:00 2001 From: Jonas Knobloch Date: Tue, 31 Mar 2026 16:21:28 +0200 Subject: Handle unknown runes --- tokenizer/bpe/tokenizer.go | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'tokenizer/bpe/tokenizer.go') diff --git a/tokenizer/bpe/tokenizer.go b/tokenizer/bpe/tokenizer.go index 1b65de2..39dd7f4 100644 --- a/tokenizer/bpe/tokenizer.go +++ b/tokenizer/bpe/tokenizer.go @@ -20,6 +20,14 @@ func (t *Tokenizer) Decode(ids []int) string { panic("unimplemented") // TODO implement } -func (t *Tokenizer) Tokenize(s string) []int { +func (t *Tokenizer) Tokenize(s string) (ids []int) { + defer func() { + if r := recover(); r != nil { + // defer UnknownRunes(t, s) + + ids = []int{} + } + }() + return t.Encode(s) } -- cgit v1.3.1