DeepSeek V4.1: A user questions who could fine-tune the base model
One post expresses frustration that DeepSeek didn't release the V4.1 base model, but predicts people would struggle even to deploy it.
TLDR
A user says DeepSeek didn't release the V4.1 base model and questions how useful a release would be: Could people fine-tune it—adapt it through further training—if deployment alone would be difficult? The post also suggests DeepSeek's base models might be the least fine-tuned historically, framing their limited use as a waste of goodwill.
Combined views
11.1K
1 Source, first seen 17d ago