mrp1mple/Llama-3.2-3B-Instruct-Football_Match_History

Hugging Face
TEXT GENERATIONPricing:Input $0.2036 / Output $1.34Concurrent Unit Cost:1Model Size:3.2BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Dec 24, 2024Architecture:Transformer0.0K Featherless Exclusive Warm

The mrp1mple/Llama-3.2-3B-Instruct-Football_Match_History model is a 3.2 billion parameter Llama 3.2-Instruct variant, fine-tuned by mrp1mple. It specializes in retrieving historical football match results for specific teams and seasons from five major European leagues (1993-94 to 2022-23). With a 32768 token context length, it excels at providing detailed team performance summaries based on its historical dataset.

Loading preview...

Model Overview

This model, mrp1mple/Llama-3.2-3B-Instruct-Football_Match_History, is a specialized instruction-tuned variant of the Llama 3.2-3B-Instruct architecture. Developed by mrp1mple, it has been fine-tuned on a comprehensive dataset of historical football match data from the top five European leagues (England, Spain, Germany, Italy, and France), covering seasons from 1993–94 to 2022–23. The model's primary function is to provide team-specific match results for a given season.

Key Capabilities

  • Historical Match Retrieval: Efficiently retrieves and presents historical match results for any specified team within the covered leagues and seasons.
  • Performance Summarization: Capable of summarizing a team's performance based on its historical data.
  • Domain-Specific Knowledge: Possesses deep knowledge of football match outcomes within its training scope.

Training Details

The model was fine-tuned using a LoRA-based approach with 4-bit precision over 1 epoch, utilizing a learning rate of 1e-4. The training data consists of over 110,000 matches sourced from a Kaggle dataset, preprocessed into an (instruction, input, output) JSONL format.

Good For

  • Historical Data Access: Ideal for users needing to quickly access past football match results.
  • Football Analytics: Can serve as a knowledge component in football analytics dashboards.
  • Research: Useful for researchers studying historical team performance or league trends.

Limitations

  • Data Scope: Knowledge is strictly limited to the specified European leagues and seasons (up to 2022–23).
  • No Predictive Capability: Not designed for predicting future match outcomes or betting odds.
  • No Real-time Updates: Will not provide information on recent or ongoing seasons beyond its training cutoff.

Finetunes
2 models