From d4975ca1433f950b86efb3ef10a0e192d657c6ee Mon Sep 17 00:00:00 2001
From: Daniel Han <danielhanchen@gmail.com>
Date: Thu, 15 May 2025 04:52:10 -0700
Subject: [PATCH] Update vision.py

---
 unsloth/models/vision.py | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/unsloth/models/vision.py b/unsloth/models/vision.py
index d57e51a820..97ed9945d7 100644
--- a/unsloth/models/vision.py
+++ b/unsloth/models/vision.py
@@ -202,10 +202,10 @@ def unsloth_base_fast_generate(
             cache_implementation = "hybrid"
     if "generation_config" in kwargs:
         kwargs["generation_config"].cache_implementation = cache_implementation
-        kwargs["generation_config"].compile_config = _compile_config
+        kwargs["generation_config"].compile_config = _compile_config if cache_implementation is not None else None
     else:
         kwargs["cache_implementation"] = cache_implementation
-        kwargs["compile_config"] = _compile_config
+        kwargs["compile_config"] = _compile_config if cache_implementation is not None else None
     pass
 
     try:
@@ -215,7 +215,6 @@ def unsloth_base_fast_generate(
         PROMPT_LOOPKUP[arch] = False
         kwargs.pop("prompt_lookup_num_tokens", None)
         with torch.inference_mode(), autocaster:
-            print(args, kwargs)
             output = self._old_generate(*args, **kwargs)
     finally:
         pass