harsh762011/numinao14

TEXT GENERATIONConcurrency Cost:1Model Size:3.8BQuant:BF16Ctx Length:32kPublished:Jan 29, 2026License:cc-by-nc-3.0Architecture:Transformer0.0K Open Weights Cold

harsh762011/numinao14 is a 3.8 billion parameter Phi-3 model developed by Harsh Srivastava, fine-tuned from unsloth/phi-4-mini-reasoning. This model was trained using Unsloth and Huggingface's TRL library, achieving 2x faster finetuning. It is designed for general language tasks, leveraging its efficient training methodology.

Loading preview...

Model Overview

harsh762011/numinao14 is a 3.8 billion parameter language model developed by Harsh Srivastava. It is a fine-tuned variant of the unsloth/phi-4-mini-reasoning base model, indicating a focus on reasoning capabilities.

Key Training Details

This model was notably trained with Unsloth and Huggingface's TRL library, which facilitated a 2x faster finetuning process. This efficient training approach allows for quicker iteration and deployment of specialized models.

Licensing

The model is released under the CC-BY-NC-3.0 license, which permits non-commercial use with attribution.

Potential Use Cases

Given its Phi-3 architecture and finetuning from a reasoning-focused base, this model is likely suitable for:

  • General text generation and understanding tasks.
  • Applications requiring efficient inference due to its optimized training.
  • Non-commercial projects that align with its license.