This is a list of frequently asked questions (FAQ) relating to the June 2024 developer contest.
For more details of the contest : Generative AI Agents Developer Contest by NVIDIA & LangChain
There is an active discussion about the contest on our developer discord : Discord
This is a living document, more FAQs may be added as we discover them.
FAQs:
Q: Which countries is this contest open to?
A: Here is the link to the list of countries that are open for the contest - Generative AI Agents Developer Contest by NVIDIA & LangChain
Q: Can one use OpenAIās API like GPT-4 turbo/o ?
A: In addition to NVIDIA and LangChain tools, you can also use OpenAIās API.
Q: Can we fine tune LLM model?
A: Yes, you can fine tune any LLM model and use it for your application. It should be noted that you will need your own compute infrastructure for fine tuning models. You can also use colab notebooks using NeMo. Few examples can be found here: GenerativeAIExamples/models/StarCoder2/lora.ipynb at main Ā· NVIDIA/GenerativeAIExamples Ā· GitHub
Q: What is NVIDIA AI Foundation endpoint?
A: The NVIDIA AI foundation endpoints are models developed by the community and further optimized by NVIDIA. You can use these models for inferencing.
Q: Is there any way to deploy self-hosted NVIDIA NIM endpoint for inferencing?
A: Self hosted NIM is not available at the moment. An alternative to that would be to deploy the model using Hugging Face
Q: Can I register using my student email address?
A: We encourage you to use your personal email address when registering for the contest.
Q: The ChatNVIDIA class from LangChain did not implement bind_tools functions and hence we get an error message when using Chat OPENAI with baseurl NIM endpoint. Is there a workaround?
A: Yes, currently tool calling is not supported and will not be available for this contest. Here(GenerativeAIExamples/notebooks/09_Agent_use_tools_leveraging_NVIDIA_AI_endpoints.ipynb at main Ā· NVIDIA/GenerativeAIExamples Ā· GitHub ) is an example you can use if you plan to use NIM endpoint. You are open to use other tools as well.
Q: Is there any requirement to be a desktop app, web app, mobile app, or could be whatever we want?
A: The goal of this Hackathon is to enable developers to create compelling applications using tools and framework offered by NVIDIA and Langchain. In addition, developers can also use other tools popular in the developer community.
Q: I donāt have the GPU cards with the spec mentioned in the contest, how do I develop my apps while taking part in the contest?
A: You can use our endpoints with deployed models. You can create an API key and get 1000 calls for free here NVIDIA AI Foundation Endpoints | š¦ļøš LangChain
Q: Can developers use multimodel LLMs in their applications?
A: Yes, developers can use multimodel LLMs in their application.
Q: How will my submission be judged?
A: Submissions will be judged based on the below parameter
Real-world application: Evaluates the impact and novelty of the project in addressing real-world challenges and the ease of use for its target audience
Technology integration: Assesses how effectively the developer has used āNVIDIAs LLM products and LangChain technologies in the project
Quality of submission: Reviews the comprehensiveness and clarity of the project details, instructions, and interesting demo
Q: Should the application be using all the tools (NVIDIA or Langchain) listed in the contest requirement?
A: No, using all the tools is not a requirement, but using one of the offerings from both NVIDIA and Langchain where applicable is important.
Q: Can participants from countries like India and China team up with other members who are participating from valid countries
A: Team entries are not allowed to participate in the contest.
Q: Should one submit the API keys for any model end points that is part of the application?
A: If you are using NVIDIAās NIM service then keep the API key section blank, for any service please provide the API key.
Recommended best practice would be to create a config file and save it in your google drive or OneDrive. Share the link to the respective drive in āquestion 15 (Link to Supplementary Materials (ex. Google Drive).
Q: How can one use NIMs from hugging face?
A: Details about how to use NIMS from hugging face can be found here: https://developer.nvidia.com/blog/nvidia-collaborates-with-hugging-face-to-simplify-generative-ai-model-deployments/
Q: What is the license of the submitted code?
A: Any open-source license should be good such as Apache 2.0, BSD or MIT.
Ask-An-Expert Session on June 14th, 2024
Some more tips were shared in the session, a recording can be viewed at:
https://developer.download.nvidia.com/forums/videos/Ask-An-Expert-Session-06-14-24.mp4