metadata

license: apache-2.0

Introducing SmallThinker-3B: A Lightweight Model Fine-tuned on QwQ Synthetic Data

This document introduces SmallThinker-3B, a new model fine-tuned from the Qwen2.5-3b-Instruct model using synthetic data generated by QwQ-32B-Preview.

Benchmark Performance

[Insert Benchmark Results Here - e.g., a table or bullet points showing performance on various metrics]

Example Metric 1: [Score/Value]
Example Metric 2: [Score/Value]
Example Metric 3: [Score/Value]

Please replace the example metrics with your actual benchmark results.

Intended Use Cases

SmallThinker is designed for the following use cases:

Edge Deployment: Its small size makes it ideal for deployment on resource-constrained devices.
Draft Model for QwQ-32B-Preview: Bamboo can serve as a fast and efficient draft model for prototyping and experimentation before utilizing the larger QwQ-32B-Preview model.

Limitations & Disclaimer

Please be aware of the following limitations:

Language Limitation: The model has only been trained on English-language datasets, hence its capabilities in other languages are still lacking.
Unpredictable Outputs: The model may produce unexpected outputs due to its size and probabilistic generation paradigm. Users should exercise caution and validate the model's responses.