Fine-Tune Gemma 4 on Your Data (LoRA Hands-On) - Open Registration
Date
Wed, May 6
Time
02:00 AM - 04:00 AM
Price
50 kr
Bring your laptop and data, and leave with a model tailored to your needs!
🔥 What to Expect
Join us for an engaging two-hour hands-on workshop dedicated to fine-tuning the Gemma 4 model, launched on March 31. This event is all about practicality - no lengthy theoretical presentations! You'll set up a training run using real H200 GPUs and walk away with a customized adapter that you can host or deploy.
🤖 Workshop Activities
- Choose your unique task or tone (think Slack voice, customer support replies, code review comments).
- Define a skill (like ticket classification, drafting briefs, or answering domain questions).
- Prepare your dataset and kick off training in Nebius AI Cloud.
- Evaluate results against the base model and witness the improvements firsthand.
🎁 What You Receive
- Access to a shared $100 pool of Nebius AI Cloud credits for training, allowing approximately 85 minutes of H200 time per attendee.
- $50 in Nebius Token Factory inference credits for testing your custom model against the base.
- A comprehensive workshop repository with data preparation scripts, training configurations, and evaluation tools.
- A functioning LoRA adapter on your chosen model: Gemma 4 E2B (5GB, laptop-friendly), E4B (8GB), or 26B-A4B (MoE).
👥 Who Should Attend
This workshop is designed for anyone who's worked with large language models (LLMs) and is eager to go beyond mere prompting. Whether you're new to model training or have experience shipping models, we’ll tailor the session to match your expertise level.
📍 Logistics
- When: Tuesday, May 5, at 7:00 PM
- Where: Frontier Tower, Floor 10 (Immersive Commons), 995 Market Street, San Francisco
- What to Bring: Your laptop, your data (or use ours), and a Hugging Face account
- Enjoy complimentary snacks and water, generously provided by Nebius!
🎙️ Hosts & Sponsors
Hosted by the Immersive Commons Applied AI team, with Lead: Rayyan Zahid and Co-speaker: Eric Mockler. Sponsored by Nebius, providing GPU credits and support.
🌍 Note: This workshop is independently organized and not endorsed by Google. Limited to 25 attendees - Doors open at 6:45 PM.
NOTE: We cannot guarantee the accuracy of the information we provide about this event. Visit the event's website to verify details such as date, opening hours, prices and location.
Frontier Tower @ 10th Floor Annex 995 Market Street, San Francisco