V4 Eyes and ears… (high-level) have a series of images that would like to train the tool to identify objects in… images update each minute, logic should run on each new image… if object(s) are found, it should notify… is this possible with v4?
Cool thought. It’s going to be very interesting (but probably very expensive) to run this every minute.
For now it’s not possible as the API is not available. May be available soon.
In regards to training. Not sure. AFAIK there hasn’t been any indication that the model will be available for further training/fine-tuning.
BUT. Starting with GPT-Vision to automate/aid training a lesser model is definitely a viable path. Especially if your goal is just to identify certain objects that have obvious defining features.
BUT. Starting with GPT-Vision to automate/aid training a lesser model is definitely a viable path. Especially if your goal is just to identify certain objects that have obvious defining features.
Thank you for your reply… yes, the lesser model is perfectly viable. “GPT-Vision” huh? Is there a date when this will be available?
Chuck
There’s no date but the rumor is at the very least it will be announced on Dev Day (Nov 6)
Thank you! This was very helpful!
Chuck