Skip to content

Fix the problem of not using the decoding method corresponding to the base model in peft mode #2865

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merrymercy merged 2 commits into from
Dec 28, 2023
Merged

Fix the problem of not using the decoding method corresponding to the base model in peft mode #2865

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
46d559e
fix the problem of not using the decoding method corresponding to the…
Jingsong-Yan Dec 27, 2023
1b03606
remove test function
Jingsong-Yan Dec 27, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Next Next commit
fix the problem of not using the decoding method corresponding to the… 
… base model in peft mode
  • Loading branch information
@Jingsong-Yan
Jingsong-Yan committed Dec 27, 2023
commit 46d559eb9cb2f73afaaf10c9e88c0936f9642a12
11 changes: 10 additions & 1 deletion fastchat/model/model_adapter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -370,10 +370,14 @@ def get_generate_stream_function(model: torch.nn.Module, model_path: str):
from fastchat.serve.inference import generate_stream

model_type = str(type(model)).lower()
is_peft = "peft" in model_type
if is_peft:
model.set_adapter(model_path)
model_type = str(type(model.base_model.model))

is_chatglm = "chatglm" in model_type
is_falcon = "rwforcausallm" in model_type
is_codet5p = "codet5p" in model_type
is_peft = "peft" in model_type
is_exllama = "exllama" in model_type
is_xft = "xft" in model_type

Expand Down Expand Up @@ -418,6 +422,11 @@ def generate_stream_peft(
else:
return generate_stream

def get_result(model, tokenizer, prompt):
for response, history in model.stream_chat(tokenizer, prompt, [], past_key_values=None, return_past_key_values=False, max_length=8192, top_p=0.8, temperature=0.95):
pass
return response


def add_model_args(parser):
parser.add_argument(
Expand Down