ai-workspace-services/litellm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(factory.py): correctly apply bos token for llama3 instruct prompt template

Browse Source
This commit is contained in:
Krrish Dholakia 2024-04-25 17:37:17 -07:00
parent 5bd50a6cc0
commit ccb864bbea
1 changed files with 2 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
litellm/llms/prompt_templates/factory.py
View File

@ -1350,11 +1350,8 @@ def prompt_factory(
return llama_2_chat_pt(messages=messages)
elif "llama3" in model and "instruct" in model:
return hf_chat_template(
model=model,
model="meta-llama/Meta-Llama-3-8B-Instruct",
messages=messages,
chat_template=known_tokenizer_config[ # type: ignore
"meta-llama/Meta-Llama-3-8B-Instruct"
]["tokenizer"]["chat_template"],
)
elif custom_llm_provider == "perplexity":
for message in messages:
@ -1369,11 +1366,8 @@ def prompt_factory(
"meta-llama/llama-3" in model or "meta-llama-3" in model
) and "instruct" in model:
return hf_chat_template(
model=model,
model="meta-llama/Meta-Llama-3-8B-Instruct",
messages=messages,
chat_template=known_tokenizer_config[ # type: ignore
"meta-llama/Meta-Llama-3-8B-Instruct"
]["tokenizer"]["chat_template"],
)
elif (
"tiiuae/falcon" in model