Model Overview

uukuguy/speechless-tora-code-7b-v1.0 is a 7 billion parameter instruction-tuned model developed by uukuguy, designed to enhance reasoning and planning abilities, particularly in coding contexts. It is fine-tuned using a comprehensive dataset of over 200,000 samples, including filtered categories from jondurbin/airoboros-2.2, Open-Orca/OpenOrca, garage-bAInd/Open-Platypus, WizardLM/WizardLM_evol_instruct_V2_196k, TokenBender/python_eval_instruct_51k, and Spider.

Key Capabilities

• Code Generation: Achieves a HumanEval Python score of 51.829, placing it competitively with larger models like CodeLlama-34B-Instruct.
• Reasoning and Planning: Enhanced through specific dataset filtering for these capabilities.
• Instruction Following: Accepts Alpaca instruction format for clear and direct task execution.

Use Cases

• Intelligent Programming Assistant: Ideal for generating code, assisting with debugging, and explaining programming concepts.
• Code-related Task Automation: Suitable for tasks requiring logical reasoning and structured output in programming domains.
• Educational Tools: Can be integrated into platforms for learning and practicing coding.