Tiny Llama 3.2 Instruct (Original)

This model is the original version of a tiny language model based on the architecture of Llama 3.2-3B-Instruct, created before any distillation or fine-tuning on specific datasets.

Model Details

  • Model Name: tiny-llama3.2-instruct (Original)
  • Architecture: Based on the Llama 3 architecture.
  • Parameters: Approximately 265 million parameters (as configured).
  • Language: English (primarily, as it's based on a pre-trained Llama model).
  • Developed By: [Your Name/Hugging Face Username]
  • License: [The license of the base Llama 3 model - likely Apache 2.0]

Intended Use

This model is intended as a small, efficient base for further experimentation, fine-tuning, or distillation. It can potentially be used for general instruction following tasks, although its capabilities may be limited compared to larger models.

Training Data

This model was created by [Here, you should describe how you created this tiny model...].

Training Procedure

The model was trained by [Describe the training procedure used to create this tiny model...].

Evaluation

[If you performed any evaluation..., describe the metrics and results here. If not, you can state that no specific evaluation was performed on this base version.]

Limitations and Potential Biases

As a smaller model based on the Llama 3 architecture, this model may have limitations...

How to Use

You can load and use this model using the transformers library...

Contact

[Your preferred contact method...]

Acknowledgements

This model is based on the Llama 3 architecture developed by Meta AI.

Downloads last month
36
Safetensors
Model size
266M params
Tensor type
FP16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support