Inspiration
Yet, the true potential of AI has been gated behind the high barriers of cost, complexity, and accessibility. We think about one question: How can we make cutting-edge AI more accessible to everyone? Inspired by the possibilities of decentralized computing and the wealth of pre-trained AI models on Hugging Face, we saw an opportunity to democratize AI deployment. Thus, GOLLM was born—a platform designed to bridge the gap between sophisticated AI models and users across the globe, no matter of their technical expertise or financial resources.
What it does
Would problem we are trying to solve?
- Decentralization: Together with GOLLM you are free in your choice of choosing model without waiting for permissions or receiving restrictions from centralized resources provider
- Cost: You can rely on market cost of resources in golem network rather than monopolized cloud market.
- Easy: If you want to test some models you just choose a name and chat. It’s nocode
How we built it
Design Phase
We started with designing a user-friendly interface that allows to deploy AI models with just a name of the model. Our goal was to ensure that anyone could use GOLLM to bring AI models to life.
Integration with Hugging Face
We developed a web page integrated with Hugging Face, enabling users to select from thousands of pre-trained models directly within GOLLM.
GOLEM NETWORK
Our platform harnesses the Golem Network for computational resources. We implemented a solution to run service across the network, allowing users to deploy and host AI models.
TELEGRAM
We created a Telegram application that allows users to interact with GOLLM just like they would chat with a friend.
Challenges we ran into
Navigating Decentralized Computing
Adapting to the decentralized nature of the Golem Network presented a learning curve. Ensuring consistent performance across a distributed network required innovative solutions and continuous optimization.
User Interface Design
Creating an interface that was both powerful and easy to use was a delicate balancing act. We iterated through multiple designs to find one that provided simplicity without sacrificing functionality.
What's next for GOLLM
Our journey with GOLLM is just beginning. As we look to the future, we're excited about expanding our platform, incorporating more AI models, and exploring new functionalities. The challenges we've faced have only strengthened our resolve to make AI deployment as accessible and straightforward as sending an email. With GOLLM, we're one step closer to that reality.
Log in or sign up for Devpost to join the conversation.