Model Overview

arcee-ai/zilo-instruct-v2-sft-filtered is a 7 billion parameter instruction-tuned model built upon the mistralai/Mistral-7B-Instruct-v0.2 architecture. It has been fine-tuned by arcee-ai using the proprietary arcee-ai/Zilo-Filtered-SQL-Instruct-v2 dataset.

Key Characteristics

• Base Model: Mistral-7B-Instruct-v0.2
• Parameter Count: 7 billion
• Context Length: 4096 tokens
• Fine-tuning Dataset: arcee-ai/Zilo-Filtered-SQL-Instruct-v2

Training Details

The model was trained with a learning rate of 2e-05 over 3 epochs. Key hyperparameters included a train_batch_size of 16, eval_batch_size of 8, and an Adam optimizer. The training process resulted in a final validation loss of 0.5474.

Potential Use Cases

Given its fine-tuning on a SQL instruction dataset, this model is likely best suited for tasks involving:

• Generating SQL queries from natural language instructions.
• Interpreting and explaining SQL commands.
• Assisting with SQL-related development and data manipulation tasks.