make gemma's output better (intel#10943)

jeanyu-habana · May 8, 2024 · 5973d6c · 5973d6c
1 parent 15ee3fd
commit 5973d6c
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/python/llm/example/GPU/HF-Transformers-AutoModels/Model/gemma/generate.py b/python/llm/example/GPU/HF-Transformers-AutoModels/Model/gemma/generate.py
@@ -48,6 +48,7 @@
                                                  load_in_4bit=True,
                                                  optimize_model=True,
                                                  trust_remote_code=True,
+                                                 mixed_precision=True,
                                                  use_cache=True)
     model = model.to('xpu')