File size: 1,972 Bytes
be16586 8084674 8971b27 be16586 8084674 8971b27 8084674 8971b27 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 |
---
emoji: 🔥
colorFrom: gray
colorTo: indigo
sdk: static
pinned: false
title: "DSTI — AI Community"
type: "organization"
tags:
- community
- education
- machine-learning
- deep-learning
- nlp
- computer-vision
- generative-ai
- open-source
- teaching
- research
library_name:
- transformers
- diffusers
- datasets
- peft
---
# DSTI — AI Community
**DSTI** is a university community focused on teaching, researching, and building with **AI**.
We share open-source models, datasets, and learning resources across **NLP**, **Computer Vision**, and **Generative AI**, with a strong emphasis on **reproducibility** and **knowledge sharing**.
> **Interests:** Deep Learning (NLP & CV), Open Source, Reproducible Research, Knowledge Sharing, MLOps, Responsible AI.
---
## 🎯 Mission & Values
- **Educate**: Provide hands-on, industry-relevant ML/AI education.
- **Open**: Build in the open—code, weights, datasets, and docs.
- **Reproduce**: Promote rigorous, repeatable experiments and clear reporting.
- **Uplift**: Help students and practitioners learn, contribute, and grow together.
---
## 🧭 Focus Areas
- **NLP**: instruction-tuned LMs, evaluation, tokenization, prompt design.
- **Computer Vision**: image classification, detection, segmentation, multimodal.
- **Generative AI**: diffusion models, LLM fine-tuning/LoRA, evaluation.
- **MLOps & Tooling**: training recipes, experiment tracking, deployment tips.
- **Responsible AI**: bias, safety, documentation, and model cards.
---
## 📦 What You’ll Find Here
- **Models**: Student projects, course baselines, research prototypes.
- **Datasets**: Curated, class-safe datasets with clear licenses and splits.
- **Spaces**: Demos, teaching assistants, benchmarking dashboards.
- **Learning Resources**: Labs, notebooks, reading lists, and cheatsheets.
> All repos aim to include a **README/model card**, **license**, **training details**, and **repro steps**.
|