Implement LLaMA (#9)

Co-authored-by: Zhuohan Li <zhuohan123@gmail.com>
2023-03-29 21:25:32 -07:00
parent a1b3de86cd
commit 80a2f812f1
7 changed files with 500 additions and 35 deletions
--- a/cacheflow/master/simple_frontend.py
+++ b/cacheflow/master/simple_frontend.py
@@ -61,4 +61,5 @@ class SimpleFrontend:
        for seq in seq_group.seqs:
            token_ids = seq.get_token_ids()
            output = self.tokenizer.decode(token_ids, skip_special_tokens=True)
+            output = output.strip()
            print(f'Seq {seq.seq_id}: {output!r}')