diff --git a/unsloth/models/loader.py b/unsloth/models/loader.py index 07396313b3..a9e5a2ca5c 100644 --- a/unsloth/models/loader.py +++ b/unsloth/models/loader.py @@ -71,7 +71,7 @@ class FastLanguageModel(FastLlamaModel): device_map = "sequential", rope_scaling = None, fix_tokenizer = True, - use_gradient_checkpointing = True, + use_gradient_checkpointing = False, *args, **kwargs, ): old_model_name = model_name