can i train the farm reader model based on the local text document ?

[Antonyjoseph4]

I have a text document and i have generated questions using the RetrieverQuestionGenerationPipeline, meanwhile i'm fetching the answer from the farm reader model using the ExtractiveQAPipeline, i just want to clarify whether if i can train the farm model with my own text document for answer generation

[ZanSara]

Have you seen this documentation page? https://docs.haystack.deepset.ai/docs/reader#fine-tuning-saving-loading-and-converting? We also have a tutorial: https://haystack.deepset.ai/tutorials/02_finetune_a_model_on_your_data

[Antonyjoseph4]

the training data given in the documents are in the form of .json format, wheras i want to train with my text data which contains paragraph format texts

[ZanSara]

readers are models trained for question answering, so you need question-answer pairs in squad forma to train such models. Have a look at our documentation on the topic: https://docs.haystack.deepset.ai/docs/domain_adaptation#user-feedback

[Antonyjoseph4]

the training data given in the documents are in the form of .json format, wheras i want to train with my text data which contains paragraph format texts