Niederbairisch chat v1
A LoRA adapter for Qwen2.5-0.5B-Instruct, trained on the CC BY 4.0 Niederbairisch dataset v0.9.3. Dataset credit: saggl. The adapter is Apache 2.0, following the base model. Training uses only the train split of translation, comprehension, and dialogue generation tasks. Validation is used for monitoring. This is an experimental style model trained on a small corpus; dialect accuracy has not been independently validated. Load with peft.PeftModel.from_pretrained(base_model, "saggl/niederbairisch-chat-v1").