>
P
Model Details
Domain:
Model Access:
Open weights (unrestricted)
Introduction
We introduce phi-1, a new large language model for code, with significantly smaller size than competing models: phi-1 is a Transformer-based model with 1.3B parameters, trained for 4 days on 8 A100s, using a selection of ``textbook quality" data from the web (6B tokens) and synthetically generated textbooks and exercises with GPT-3.5 (1B tokens). Despite this small scale, phi-1 attains pass@1 accuracy 50.6% on HumanEval and 55.5% on MBPP. It also displays surprising emergent properties compared to phi-1-base, our model before our finetuning stage on a dataset of coding exercises, and phi-1-small, a smaller model with 350M parameters trained with the same pipeline as phi-1 that still achieves 45% on HumanEval.
Benchmarking
FLOPs332000000000000000000
Notes: 6ND = 6 *1.3*10^9 parameters * 51 * 10^9 tokens = 3.978e+20 312000000000000 FLOP/s * 8 GPUs *103 hours *3600 sec / hour *0.3 [assumed utilization] = 2.7765504e+20 geometric mean sqrt(2.7765504e+20*3.978e+20) = 3.3234195e+20
Training
Training Code Accessibilityhttps://huggingface.co/microsoft/phi-1 MIT license
HardwareNVIDIA A100
Hardware Quantity8
Size Notes: A filtered code-language dataset, which is a subset of The Stack and StackOverflow, obtained by using a language model-based classifier (consisting of about 6B tokens). • A synthetic textbook dataset consisting of <1B tokens of GPT-3.5 generated Python textbooks. • A small synthetic exercises dataset consisting of ∼180M tokens of Python exercises and solutions For the 1.3B models, phi-1 and phi-1-base are checkpoints after training on 51B tokens (770 GPU hours) Training tokens: 54B tokens (7B unique tokens)
Parameters
Parameters1300000000
Notes: 1.3B The architecture for our 1.3B parameter phi-1 model consists of 24 layers, hidden dimension of 2048, MLP-inner dimension of 8192, and 32 attention heads of dimension 64 each.
Authors
Suriya Gunasekar, Yi Zhang, Jyoti Aneja, Caio César Teodoro Mendes, Allie Del Giorno, Sivakanth Gopi, Mojan Javaheripi, Piero Kauffmann, Gustavo de Rosa, Olli Saarikivi, Adil Salim, Shital Shah, Harkirat Singh Behl, Xin Wang, Sébastien Bubeck, Ronen Eldan, Adam Tauman Kalai, Yin Tat Lee, Yuanzhi Li
Biology