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
279 Jun 26, 2022
47 Dec 30, 2022
75 Jul 15, 2022
150 Dec 15, 2022
1 Jun 02, 2022
6 Sep 21, 2022
7.9k Jan 06, 2023
132 Sep 23, 2022
242 Dec 12, 2022
298 Jan 02, 2023
7 Nov 02, 2022
54 Dec 08, 2022
1 Jan 12, 2022
130 Dec 26, 2022
59 Dec 28, 2022
24 Oct 24, 2022
0 Aug 17, 2021
118 Dec 12, 2022
367 Dec 22, 2022
20 Aug 04, 2022