EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Super tiny typo fix (#2564)

Browse Source
This commit is contained in:
fzyzcjy
2024-12-27 00:28:01 +08:00
committed by GitHub
parent ed91e003bb
commit 44f011d224
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/models/llama.py
View File

@@ -325,8 +325,8 @@ class LlamaForCausalLM(nn.Module):
self.config = config
self.quant_config = quant_config
self.model = LlamaModel(config, quant_config=quant_config)
# Llama 3.2 1B Insturct set tie_word_embeddings to True
# Llama 3.1 8B Insturct set tie_word_embeddings to False
# Llama 3.2 1B Instruct set tie_word_embeddings to True
# Llama 3.1 8B Instruct set tie_word_embeddings to False
if self.config.tie_word_embeddings:
self.lm_head = self.model.embed_tokens
else: