TeichAI/Qwen3-1.7B-Gemini-2.5-Flash-Lite-Preview-Distill

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:Nov 12, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

TeichAI/Qwen3-1.7B-Gemini-2.5-Flash-Lite-Preview-Distill is a 2 billion parameter language model developed by TeichAI, fine-tuned from unsloth/Qwen3-1.7B-unsloth-bnb-4bit. It was trained on 1000 examples from Gemini 2.5 Flash Lite Preview 09-2025, leveraging Unsloth and Huggingface's TRL library for 2x faster training. This model is optimized for tasks requiring efficient processing within a 40960 token context length, making it suitable for applications benefiting from distilled knowledge from a larger, more capable model.

Loading preview...

Model Overview

TeichAI/Qwen3-1.7B-Gemini-2.5-Flash-Lite-Preview-Distill is a 2 billion parameter language model developed by TeichAI. It is fine-tuned from unsloth/Qwen3-1.7B-unsloth-bnb-4bit and utilizes a substantial 40960 token context length.

Key Characteristics

  • Distilled Knowledge: This model was trained on 1000 examples sourced directly from the Gemini 2.5 Flash Lite Preview 09-2025 model, aiming to distill capabilities from a more advanced source.
  • Efficient Training: Development leveraged Unsloth and Huggingface's TRL library, resulting in a 2x faster training process compared to conventional methods.
  • Qwen3 Architecture: Built upon the Qwen3 architecture, providing a solid foundation for language understanding and generation tasks.

Use Cases

This model is particularly well-suited for applications where:

  • Resource Efficiency is Key: Its 2 billion parameter size makes it more efficient for deployment on constrained hardware.
  • Knowledge Distillation is Beneficial: Leveraging insights from a powerful model like Gemini 2.5 Flash Lite Preview, it can perform tasks that benefit from this distilled intelligence.
  • Fast Prototyping and Development: The efficient training methodology suggests it can be quickly adapted or fine-tuned for specific downstream tasks.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p