GPT - gMLP
This repository will attempt to crack long context autoregressive language modeling (GPT) using variations of gMLPs. Specifically, it will contain a variant that does gMLP for local sliding windows. The hope is to be able to stretch a single GPU to be able to train context lengths of 4096 and above efficiently and well.
GPT is technically a misnomer now, since there will be no attention (transformer) at all contained in the architecture.
Install
$ pip install g-mlp-gpt
Usage
import torch
from g_mlp_gpt import gMLPGPT
model = gMLPGPT(
num_tokens = 20000,
dim = 512,
depth = 4,
seq_len = 1024,
window_size = (128, 256, 512, 1024) # window sizes for each depth
)
x = torch.randint(0, 20000, (1, 1000))
logits = model(x) # (1, 1000, 20000)
16k context length
import torch
from g_mlp_gpt import gMLPGPT
model = gMLPGPT(
num_tokens = 20000,
dim = 512,
seq_len = 16384,
depth = 8,
reversible = True,
window = (128, 128, 256, 512, 1024, 1024, 2048, 2048, 4096, 4096, 8192, 8192),
axial = (1, 1, 1, 1, 1, 1, 2, 2, 4, 4, 8, 8)
).cuda()
x = torch.randint(0, 20000, (1, 16384)).cuda()
logits = model(x) # (1, 16384, 20000)
Citations
@misc{liu2021pay,
title = {Pay Attention to MLPs},
author = {Hanxiao Liu and Zihang Dai and David R. So and Quoc V. Le},
year = {2021},
eprint = {2105.08050},
archivePrefix = {arXiv},
primaryClass = {cs.LG}
}
9 Nov 16, 2022
24 Oct 26, 2022
331 Jan 3, 2023
74 Dec 5, 2022
28 Feb 20, 2022
81 Nov 26, 2022
39 Nov 30, 2022
49 Dec 22, 2022
120 Dec 12, 2022
96 Nov 05, 2022
7 May 23, 2019
1 Feb 04, 2022
71 Jan 06, 2023
234 Dec 07, 2022
335 Dec 22, 2022
113 Dec 29, 2022
209 Jan 05, 2023
150 Dec 26, 2022
530 Dec 27, 2022
161 Jan 03, 2023
5 Jul 27, 2022
166 Dec 28, 2022
18 Nov 06, 2022
63 Dec 16, 2022
60 Nov 14, 2022
545 Dec 28, 2022
5 Oct 12, 2022
20 Sep 04, 2022