Phi-1

Training compute

3.3×10²⁰ FLOP

Parameters

1.3B

Published

Oct 2, 2023

Phi-1 is an AI model developed by Microsoft Research (United States), first published in October 2023. It works in the language domain, on tasks such as language modeling/generation and code generation.

Training it took an estimated 3.3×10²⁰ FLOP of compute (estimation method: operation counting,hardware). The model has 1,300,000,000 parameters. Training ran on 8 NVIDIA A100 for about 103 hours.

Access: Open weights (unrestricted). Its weights are openly available. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Microsoft Research
Country of organization: United States
Domain: Language
Task: Language modeling/generation, Code generation
Training compute: 3.3×10²⁰ FLOP
Compute estimation method: Operation counting, Hardware
Parameters: 1,300,000,000
Training hardware: NVIDIA A100
Chips used: 8
Training time: 103 h
Training power draw: 6.4 kW
Model accessibility: Open weights (unrestricted)
Open weights: Yes
Epoch confidence: Confident