Instructions to use NemoStation/Marlin-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use NemoStation/Marlin-2B with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForCausalLM processor = AutoProcessor.from_pretrained("NemoStation/Marlin-2B", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("NemoStation/Marlin-2B", trust_remote_code=True) - Notebooks
- Google Colab
- Kaggle
👉 Marlin-2B model? "Open Source" -- my ass!
@HappyyPablo Read the fine print before you celebrate your own "open sourcing" lie:
👉 Marlin-2B model? "Open Source" -- my ass!
https://hyperscope.link/1/2/4/9/9/7/Marlin-2B-model-Open-Source-my-ass-124997.html
You’re not open sourcing anything. You’re dropping a Business Source License 1.1 that explicitly bans anyone from using Marlin-2B in a way that competes with NemoStation’s paid version. That’s not open source — that’s source-available with a leash. The Open Source Definition requires no discrimination against fields of endeavor (criterion #6) and no royalty fees (criterion #1). Your license violates both. The GNU Four Freedoms? You’ve gutted freedoms 0, 2, and 3 for the first two years. But you knew that, didn’t you?
You’re counting on people not reading the license. You’re trading on decades of community trust in the term "open source" to farm retweets, followers, and free labor — while secretly reserving the right to sue anyone who actually builds a successful business on your model. That’s not open. That’s not source. That’s deceptive marketing, plain and simple.
The article calls it exactly what it is: you wanted the benefits of open source without granting the freedoms. Next time, just say "source available under BSL 1.1" and let people decide. But you won’t, because that doesn’t get you the clout. Shameful.