From 07228828a0a66ea475fa324ecdc3a0ffba872641 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Sat, 20 Jul 2024 11:53:32 -0700 Subject: [PATCH] Update llama.py --- unsloth/models/llama.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py index ca4e651598..32610bbfda 100644 --- a/unsloth/models/llama.py +++ b/unsloth/models/llama.py @@ -1068,6 +1068,8 @@ def _wrap_fast_inference(generate, device_type, dtype, model): # For newer HF kwargs["cache_implementation"] = "dynamic" + print(kwargs) + # Set pad token # old_pad_token_id = getattr(model.config, "pad_token_id", None) # old_eos_token_id = getattr(model.config, "eos_token_id", None)