We are excited to welcome you to the second episode of the Cohere AI Hackathon! In partnership with Cohere, we are giving lablab.ai community access to one of the most powerful AI language models on the market. Join us and innovate to build a better future for everyone!
Our AI hackathon brought together a diverse group of participants, who collaborated to develop a variety of impressive projects based on:
1031
Participants
54
Teams
9
AI Applications
Web application that allows user to find and browse books based on short prompts, keywords or short story plots.
PowerPuff
From a topic, generate a list of subtopics and explore them with a visual image for each. The subtopics are presented in a manageable way by showing only the most related subtopics at any given time. This naturally leads to exploring associatively like our brains are structured, from broad to deep. Each topic has a handy explanation in text from cohere generate, too!
ThePathForward
IntelliChat is an app built specifically for Chat based and Conversational AI projects. *Who is this for?* - Businesses and Enterprises who are looking to optimize and enhance their Conversational AI Workflow.
💬IntelliChat
This event has now ended, but you can still register for upcoming events on lablab.ai. We look forward to seeing you at the next one!
Checkout Upcoming Events →Submissions from the teams participating in the Cohere AI Hackathon #2 event and making it to the end 👊
The app lets you classify your text into Elementary, Intermediate and Advanced according to the level of complexity. Simplify functionality lets you generate an alternate simplified version of it.
AKR
Tarinyoom is a Discord chat bot that allows users to search for messages using a fuzzy search to better retrieve past messages. The bot uses the co:here API "Embed" tool to map discord messages onto their corresponding embedding vectors. These vectors are then stored in a back-end pinecone.io instance, so that their cosine similarities can be evaluated. The bot has two main commands: "/tsleuth" and "/tsearch <string>". "/tsleuth" instructs the bot to crawl the server that the command was issued in, storing the vector embeddings of any messages found as entries in the pinecone.io instance. "/tsearch <string>" instructs the bot to find the message whose vector embedding most closely resembles the vector embedding of the passed parameter, as evaluated by pinecone.io. Using these two commands, users can search for a past conversation with only a high-level recollection of the topics discussed, instead of having to exactly match to text in that conversation, as required by the current Discord search feature.
Tarinyoom
From a topic, generate a list of subtopics and explore them with a visual image for each. The subtopics are presented in a manageable way by showing only the most related subtopics at any given time. This naturally leads to exploring associatively like our brains are structured, from broad to deep. Each topic has a handy explanation in text from cohere generate, too!
ThePathForward
IntelliChat is an app built specifically for Chat based and Conversational AI projects. *Who is this for?* - Businesses and Enterprises who are looking to optimize and enhance their Conversational AI Workflow.
💬IntelliChat
A travel recommendation for users of any kind. Users are asked a series of questions related to their personality. Then, based on those responses, the site returns a location most relevant to their personal interests.
Googol
We all understand what Cohere Generate and Classify do. But embed ...?! Introducing ... Cohere Analyze. You can upload any csv file. You then have 3 options. EDA gives you an overview of the data and you can get some general exploratory data analysis. Cluster, allows you to do some group analysis, with keywords generated from the body of the text and using the titles. And finally, Search, allows you to query the data and retrieve the closest match. This has significant business value because now you can gain business insights using cluster that are non-trivial. You can search a knowledge-base or other internal data sources for information. Cohere Analyze is what we have all been waiting for, to better utilize Embed.
Innovate
This app is the combination of mutiples work that can be done using LLM. We tried to make the llm to do things by changing the prompt . There is a chatbot. We have text summarizer, that can make good summarization. We also implemented machine translation, we have 6 languages that you can choose from. There other things we have are code explanation and story generation. It is created using streamlit library. A free python library to run ml apps as web service.
DreamDream
Constantly changing and evolving solutions, services and products available in the telecom, mobility, and IT management industry create sourcing, evaluation, and selection complexities that SMBs (small and mid-sized businesses) are not always equipped to address due to limited time or resources. How do SMBs more easily and efficiently sort out the thousands of options available to them? Our solution is an IT product Service Recommender System which will help SMBs find verified solutions to simplify the sourcing, evaluation, and selection of telecom, mobility, and IT services and solutions. Input: SLA agreements Output: Classify the SLA agreements by Subscription and One-off
Lord Christ
Web application that allows user to find and browse books based on short prompts, keywords or short story plots.
PowerPuff