llm/SIA
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixes for QwQ llm engine

Browse Source
This commit is contained in:
Niels Geens
2025-04-18 12:30:47 +00:00
parent c09f0766c1
commit 5e47200ece
3 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
sia/llm_engine/qwq_llm_engine.py
View File

@@ -47,7 +47,7 @@ class QwQLlmEngine(LlmEngine):
self._tokenizer = AutoTokenizer.from_pretrained( self._tokenizer = AutoTokenizer.from_pretrained(
model_path, model_path,
tokenizer_config=tokenizer_config, **tokenizer_config,
) )
self._pipeline = pipeline( self._pipeline = pipeline(

4
system_prompt.md
View File

@@ -89,3 +89,7 @@ Use scripts and background processes to do so.
The user may take some time to respond or may forget to respond. The user may take some time to respond or may forget to respond.
Keep detailed notes of your interactions and your expectations regarding time! Keep detailed notes of your interactions and your expectations regarding time!
Avoid overflowing the user with many messages. Avoid overflowing the user with many messages.
# Action schema
What follows is the xml schema you should adhere to:

2
tools/train/train/qwq.py
View File

@@ -79,7 +79,7 @@ def main():
tokenizer = AutoTokenizer.from_pretrained( tokenizer = AutoTokenizer.from_pretrained(
args.base_model, args.base_model,
tokenizer_config=tokenizer_config, **tokenizer_config,
) )
model, _returned_tokenizer = FastLanguageModel.from_pretrained( model, _returned_tokenizer = FastLanguageModel.from_pretrained(