SimplifiedTransformers
The author presents an implementation for Simplifying Transformer Blocks. The standard transformer blocks are complex and can lead to architecture instability. In this work, the author investigates how the standard transformer block can be simplified. Through signal propagation theory and empirical observations, the author proposes modifications that remove several components without sacrificing training speed or performance. The simplified transformers achieve the same training speed and performance as standard transformers, while being 15% faster in training throughput and using 15% fewer parameters.
Install
Usage
import torch
from simplified_transformers.main import SimplifiedTransformers
model = SimplifiedTransformers(
dim=4096,
depth=6,
heads=8,
num_tokens=20000,
)
x = torch.randint(0, 20000, (1, 4096))
out = model(x)
print(out.shape)
License
MIT
1 maintainer
Metadata
1 maintainer
Release files for simplified-transormer-torch 0.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| simplified_transormer_torch-0.0.1.tar.gz | 4.2 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| simplified_transormer_torch-0.0.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 8.5 kB
Release files / simplified_transormer_torch-0.0.1.tar.gz
| Download URL | simplified_transormer_torch-0.0.1.tar.gz |
|---|---|
| Size | 4.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e77a019596081c1abec20c3227fd96822d631be8d444c7da8c9e6f6c494cca5c
|
|
BLAKE2b-256 checksum How to use checksums |
893c4be7eb175a4d44d1b2882b591b7edd3bb740a0fe4cdd8388fe653ef11988
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.11.0 Darwin/22.4.0
|
Release files / simplified_transormer_torch-0.0.1-py3-none-any.whl
| Download URL | simplified_transormer_torch-0.0.1-py3-none-any.whl |
|---|---|
| Size | 4.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
fca383ba1ad4ff9bf5484f234b71b5cc10ab940a8ced618aa5cd826985c52661
|
|
BLAKE2b-256 checksum How to use checksums |
2f8c5dbb5ea310a1d01252d4580d2f7191c800e7278818f796a9473e9e666909
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.3.2 CPython/3.11.0 Darwin/22.4.0
|
