After the session
Where to go next
Today you ran a model on your own laptop, talked to it from Python, measured it against the cloud, and gave it tools. What comes next depends on how you want to learn.
Keep measuring, with your own tasks
Your local-model-workshop folder is already a small lab. Three ways to keep using it:
- Add your own tests to
measure.py. Take a real request from your work, swap its data for sample data, and compare it locally and in the cloud. That measurement is worth more than any opinion. - Try another model. If you have 16 GB of RAM or more, download a bigger one, like
qwen/qwen3.5-9b(about a 6 GB download and about 7 GB of memory), change themodelfor"local"inproviders.py, and run the measurement again. You'll see with your own numbers how much you gain in correct answers and how much you lose in speed. - Give
agent.pya new tool. For example, one that adds up all the categories, so the grand total no longer depends on the model knowing how to add.
And when you're done, free up the memory: lms unload --all and lms server stop.
Today: your first local model
|
+----------------+----------------+
| |
v v
At your own pace, on your own Guided, with projects
NIEVA ecosystems NIEVA bootcampsAt your own pace: NIEVA's ecosystems
Ecosystems are NIEVA's learning paths: courses and guides organized by topic, so you know what to learn first and what next. Go in, pick the topic you care about, and move at your own pace.
Guided and professional: the bootcamps
If you want something more structured, with a group, dates, deliverables, and reviews of what you build, we have the bootcamps. It's the path to go from trying out models to seriously building AI systems.
Your certificate of attendance
If you create your account on the platform and mark all the lessons of this workshop's material as completed, you get a Certificate of attendance automatically.
Before that, check your profile: there you can set your full name for certificates, which is the one that appears on the certificate.
Thank you
Thanks for coming and for giving your first model a try. See you soon.