QuestionAnswering
Hello, I don't know if the question is relevant here. But does this model have the ability to answer in the format transformers.pipeline ("question-answering", ...), that is, when I have a context and a question, and I need to get an answer from the model based on my context.
Hello! Yes, this model is capable of handling question-answering tasks effectively. For a specific setup where you have a context and a question and need the model to generate an answer based on the context, I recommend using the model available at https://huggingface.co/DeepMount00/Mistral-RAG. This model is designed specifically for the question-answering task.
Alternatively, you could use LLaMA-3, but you would need to construct an appropriate prompt tailored to the question-answering task. Both options are suitable, but choosing the right one depends on your specific needs and constraints.
just to clarify: For both the Mistral-RAG model and LLaMA-3, it's important not to use the transformers.pipeline("question-answering", ...)