<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

mistral-docbot

This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.1-GPTQ

Model Card Description

Model Description

The model is designed to generate a line of treatment and diagnosis in response to a given prompt of symptoms. It essentially serves as a tool to assist in medical scenario simulations or provide initial guidance based on presented symptoms.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• training_steps: 500
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1

Prompt Template

### Human: If you are a doctor, please answer the medical questions based on the patient's description.

### Patient's Description: {symptoms}
### Assistant: """"