CompassioninMachineLearning/pretrainingBasellama3kv3

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Jan 19, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

CompassioninMachineLearning/pretrainingBasellama3kv3 is an 8 billion parameter Llama-based language model developed by CompassioninMachineLearning. This model was pre-trained and optimized for efficiency, utilizing Unsloth and Huggingface's TRL library for 2x faster training. It offers a 32768 token context length, making it suitable for applications requiring processing of longer sequences.

Loading preview...

Model Overview

CompassioninMachineLearning/pretrainingBasellama3kv3 is an 8 billion parameter Llama-based language model developed by CompassioninMachineLearning. This model is notable for its efficient pre-training process, which leveraged Unsloth and Huggingface's TRL library to achieve a 2x speedup in training.

Key Characteristics

  • Architecture: Llama-based model.
  • Parameter Count: 8 billion parameters.
  • Context Length: Supports a 32768 token context window.
  • Training Efficiency: Pre-trained with Unsloth and Huggingface's TRL library for accelerated development.
  • License: Released under the Apache-2.0 license.

Potential Use Cases

This model is well-suited for developers and researchers looking for an efficiently trained Llama-based model with a substantial context window. Its optimized training process suggests potential benefits in scenarios where rapid iteration or resource-conscious deployment is important.