Reformer
a Pytorch implementation of the Reformer Network (https://openreview.net/pdf?id=rkgNKkHtvB)
Much of this code base is loosely translated from the jax implementation found here from Google: https://github.com/google/trax/blob/master/trax/models/research/reformer.py
How to use
All of the hard work has been taken care of, all you need to do is instantiate the model!
from reformer_lm.reformer_lm import ReformerLM
import torch
test = torch.rand((4, 4, 64))
model = ReformerLM(
vocab_size=300000,
d_in=test.shape[-2],
d_out=test.shape[-1],
n_layers=6,
n_heads=1,
attn_k=test.shape[-1],
attn_v=test.shape[-1],
)
output = model(test)
print(output)
This model is still in testing, and will therefore continue to see updates. PRs are welcomed! Feel free to take advantage of the Docker container for development. I have been working in notebooks to test code with the original paper, and then I refactor my code back into the package
1 maintainer
Metadata
1 maintainer
Release files for reformer-lm 1.0.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| reformer_lm-1.0.1.tar.gz | 5.6 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| reformer_lm-1.0.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 13.3 kB
Release files / reformer_lm-1.0.1.tar.gz
| Download URL | reformer_lm-1.0.1.tar.gz |
|---|---|
| Size | 5.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e607623fc53b06d2ad459c467a63fa2708e5e0e05aa7909adf309a6347820f37
|
|
BLAKE2b-256 checksum How to use checksums |
2928557c9269d29bd780fbda91e46d39ae03ab30b38c8be2f8f9610a278b8255
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/3.1.1 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.41.1 CPython/3.8.0
|
Release files / reformer_lm-1.0.1-py3-none-any.whl
| Download URL | reformer_lm-1.0.1-py3-none-any.whl |
|---|---|
| Size | 7.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
7cae226bb9eb07e74dbf8b643b595fadf6bf6166f727f4fc46f1d4929892bb90
|
|
BLAKE2b-256 checksum How to use checksums |
7ad375ecb2cf438a06bd8031dd427427e46f6229df37b262fe9dbdcf6bff454c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/3.1.1 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.41.1 CPython/3.8.0
|
Security for reformer-lm 1.0.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
