| | --- |
| | title: README |
| | emoji: 🐢 |
| | colorFrom: purple |
| | colorTo: pink |
| | sdk: static |
| | pinned: false |
| | --- |
| | |
| | ## Pandalla.ai |
| |
|
| | <p align="center"> |
| | <img src="https://raw.githubusercontent.com/pandalla/pandalla.ai/refs/heads/main/public/images/logo/pandalla.svg" width="650" style="margin-bottom: 0.2;"/> |
| | <p> |
| | <h5 align="center"> Grow Together ⭐ </h5> |
| | <h4 align="center"> [<a href="https://github.com/pandalla/pandalla.ai">GitHub</a> | <a href="https://pandalla.ai/">DataTager</a>]</h4> |
| | |
| | **Long-term Focus:** |
| | |
| | - Our company is dedicated to long-term specialization in **synthetic data**, **metaphysics**, and **psychology LLM**, exploring how these fields can intersect with AI. |
| |
|
| | **Product:** DataTager |
| |
|
| | **Website:** [Pandalla.AI](https://pandalla.ai/) |
| |
|
| | - **Description:** DataTager is a tool designed to evaluate and generate the training data needed for large language models. We believe it's more important for individuals and enterprises to fine-tune large models easily and create models tailored to their specific business needs, rather than just choosing models with the highest benchmarks. |
| |
|
| | **Philosophy:** |
| |
|
| | - We published a paper titled "AnyTaskTune," advocating that **Task Fine-Tuning** based on real-world scenarios is crucial. This approach is more significant than using universally high-scoring models. |
| |
|
| | **Resources:** |
| |
|
| | - We have open-sourced various subtask datasets across multiple domains to support the community. These resources are available on our website for anyone interested in specific task fine-tuning. |
| |
|
| | Explore more on how to fine-tune your tasks efficiently with our resources at [Pandalla.AI](https://pandalla.ai/). |
| |
|
| |
|
| |
|
