Yi-VL-34B

Training compute

1.9×10²² FLOP

Parameters

34B

Published

Jan 23, 2024

Yi-VL-34B is an AI model developed by 01.AI (China), first published in January 2024. It works in the vision, language and multimodal domain, on tasks such as visual question answering and language modeling/generation.

Training it took an estimated 1.9×10²² FLOP of compute (estimation method: hardware,operation counting). The model has 34,000,000,000 parameters. Training ran on 128 NVIDIA H100 SXM5 80GB for about 240 hours.

Access: Open weights (restricted use). Its weights are openly available. It is built on top of Yi-34B,CLIP ViT-H/14 - LAION-2B. Epoch AI rates the confidence of this record as confident.

Full record

Organization: 01.AI
Country of organization: China
Domain: Vision, Language, Multimodal
Task: Visual question answering, Language modeling/generation
Training compute: 1.9×10²² FLOP
Compute estimation method: Hardware, Operation counting
Parameters: 34,000,000,000
Training hardware: NVIDIA H100 SXM5 80GB
Chips used: 128
Training time: 240 h
Training power draw: 177.5 kW
Model accessibility: Open weights (restricted use)
Open weights: Yes
Base model: Yi-34B, CLIP ViT-H/14 - LAION-2B
Epoch confidence: Confident