diff --git a/unsloth/models/llama.py b/unsloth/models/llama.py index e3de5f6347..9746bcb885 100644 --- a/unsloth/models/llama.py +++ b/unsloth/models/llama.py @@ -1890,8 +1890,8 @@ class FastLlamaModel: gc.collect() torch.cuda.empty_cache()""" multiplier = \ - "4.5 if getattr(model.config, 'quantization_config', {'load_in_4bit' : False})['load_in_4bit'] else "\ - "8.0 if getattr(model.config, 'quantization_config', {'load_in_8bit' : False})['load_in_8bit'] else 1.0" + "4.5 if getattr(model.config, 'quantization_config', \\{'load_in_4bit' : False\\})['load_in_4bit'] else "\ + "8.0 if getattr(model.config, 'quantization_config', \\{'load_in_8bit' : False\\})['load_in_8bit'] else 1.0" debug_info = debug_info.replace("multiplier__", "(" + multiplier + ")") debug_info = debug_info.replace("P__", "get_model_param_count")