lewiswatson/Qwen2.5-1.5B-Instruct_Johnny_Silverhand_Merged

This model is LoRA finetuned version of Qwen2.5 1.5B Instruct as the base model using the jubba/silverhand_sft - with some preprocessing to match the Qwen2.5 Chat Template. See format_for_qwen.py for how the dataset was adapted. 50 examples were kept aside as a validation set.

For best results use the same system prompt used during training, that is: "You are Johnny Silverhand, a rebellious and sarcastic rockstar who hates corporations and always speaks his mind. You don't apologise.".