From 79146518b6ea6acb21429129d0feda372e247e89 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Tue, 4 Mar 2025 19:39:49 -0800 Subject: [PATCH] Update llama.py --- unsloth/models/llama.py | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py index fa37fc34bf..2155fff047 100644 --- a/unsloth/models/llama.py +++ b/unsloth/models/llama.py @@ -1554,10 +1554,10 @@ def unsloth_fast_generate( pass # Must patch accelerate for Xformers - if accelerate_new_send_to_device is not None: - import accelerate.utils.operations - accelerate.utils.operations.send_to_device = accelerate_new_send_to_device - pass + # if accelerate_new_send_to_device is not None: + # import accelerate.utils.operations + # accelerate.utils.operations.send_to_device = accelerate_new_send_to_device + # pass # For newer HF kwargs["cache_implementation"] = "dynamic" @@ -1580,9 +1580,9 @@ def unsloth_fast_generate( pass # Return accelerate back - if accelerate_new_send_to_device is not None: - accelerate.utils.operations.send_to_device = accelerate_old_send_to_device - pass + # if accelerate_new_send_to_device is not None: + # accelerate.utils.operations.send_to_device = accelerate_old_send_to_device + # pass FastLlamaModel.for_training(self)