abcorrea/llama-3.2-1b-tinystories-ft-25k

Hugging Face
TEXT GENERATIONPricing:Input $0.108 / Output $0.804Concurrent Unit Cost:1Model Size:1BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Apr 2, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Warm

abcorrea/llama-3.2-1b-tinystories-ft-25k is a 1 billion parameter Llama-based language model developed by abcorrea. This model was fine-tuned from abcorrea/llama-3.2-1b-wiki-ft-v1 and optimized for training speed using Unsloth and Hugging Face's TRL library. It is designed for efficient deployment and tasks requiring a compact yet capable language model. Its primary differentiator is its accelerated training methodology, making it suitable for rapid iteration and resource-constrained environments.

Loading preview...

Model Overview

abcorrea/llama-3.2-1b-tinystories-ft-25k is a compact 1 billion parameter language model based on the Llama architecture, developed by abcorrea. This model represents a further fine-tuning of the abcorrea/llama-3.2-1b-wiki-ft-v1 base model. A key aspect of its development is the utilization of Unsloth and Hugging Face's TRL library, which enabled a 2x faster training process.

Key Characteristics

  • Architecture: Llama-based, providing a familiar and robust foundation.
  • Parameter Count: 1 billion parameters, offering a balance between capability and computational efficiency.
  • Training Efficiency: Leverages Unsloth for significantly accelerated fine-tuning, reducing training time by half.
  • Origin: Fine-tuned from abcorrea/llama-3.2-1b-wiki-ft-v1, indicating a specialized training lineage.
  • License: Distributed under the Apache-2.0 license, allowing for broad usage and modification.

Use Cases

This model is particularly well-suited for scenarios where:

  • Rapid Prototyping: The accelerated training makes it ideal for quick experimentation and iteration.
  • Resource-Constrained Environments: Its smaller size (1B parameters) allows for deployment on hardware with limited computational resources.
  • Specific Domain Adaptation: As a fine-tuned model, it can be further adapted to niche datasets or tasks efficiently.
  • Educational and Research Purposes: Its manageable size and efficient training make it accessible for learning and exploring LLM fine-tuning techniques.

Finetunes
1 models