Yandex
80B Model Post-Trained on 3 GPUs—Dev Calls V1 'Basically Useless'
Developer post-trains Yandex's 80B model on 3 V100 GPUs, calls v1 'basically useless.' Training barriers are falling fast, but the runnable-to-usable
Back to home
AliceAI
3 articles tagged with this topic
Yandex
Reddit User Trains Yandex 80B Model at Home — AI Fine-Tuning Is Now a One-Person Job
A Reddit user is publicly fine-tuning Yandex's 80B Foundation model at home, 40% complete. The signal: enterprise-grade AI customization no longer req
AliceAI
One Dev Fine-Tunes 80B Model on Three V100s: Post-Training Cost Curve Drops
Reddit user jjusko20 fine-tunes an 80B open-source model on three 2017 V100s. The threshold for LLM post-training is dropping to individual tinkerers.