Usama1100/humanizer-llama3-8b-fp16-v2

TEXT GENERATIONPricing:Input $0.37 / Cached $0.074 / Output $0.38Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Aug 20, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

Usama1100/humanizer-llama3-8b-fp16-v2 is an 8 billion parameter Llama 3 model developed by Usama1100, fine-tuned from unsloth/llama-3-8b-bnb-4bit. This model was trained using Unsloth and Huggingface's TRL library, enabling 2x faster fine-tuning. It is designed for applications requiring an efficient and optimized Llama 3 base, leveraging its 8192 token context length.

Loading preview...

Model Overview

Usama1100/humanizer-llama3-8b-fp16-v2 is an 8 billion parameter Llama 3 model, fine-tuned by Usama1100. It is based on the unsloth/llama-3-8b-bnb-4bit architecture and utilizes the Unsloth library in conjunction with Huggingface's TRL library for accelerated training.

Key Characteristics

  • Base Model: Fine-tuned from Llama 3 8B.
  • Training Efficiency: Leverages Unsloth for 2x faster fine-tuning, making it an efficient choice for custom applications.
  • License: Distributed under the Apache-2.0 license.

Use Cases

This model is suitable for developers looking for an optimized Llama 3 8B variant that benefits from faster training methodologies. Its foundation on Llama 3 suggests applicability across a range of general-purpose language tasks, with potential for further specialization through additional fine-tuning.