A library for fast, smart transformers.
Project description
What is this?
This is a collection of layers designed to achieve state-of-the-art results in the field of natural language processing. It consists of a collection of layers, loss functions, and utilities which either implement recent transformer discoveries, or which have no analogue in the current NLP literature. It is built in torch.
The library primarily forcuses around working with the word embeddings.
It has the following desirable properties when used correctly
- torchscript compatible
- O(N) with respect to words input (configured correctly)
- Ensembles natively supported.
Documentation
See https://torch-supertransformerlib.readthedocs.io/en/latest/ for documentation
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file supertransformerlib-2.0.post0.tar.gz
.
File metadata
- Download URL: supertransformerlib-2.0.post0.tar.gz
- Upload date:
- Size: 16.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.1 CPython/3.9.14
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | e24b89cfb807be1e3ed91a54715af5115b0639951cc71488f2f2359d725cc2ad |
|
MD5 | 1dc5fa71b9ebc56dcb575ce054d1c591 |
|
BLAKE2b-256 | ec5356350e131f03c7c1e83c2619b393ee2acea48309cb59cfc8593cfd875dfb |
Provenance
File details
Details for the file supertransformerlib-2.0.post0-py2.py3-none-any.whl
.
File metadata
- Download URL: supertransformerlib-2.0.post0-py2.py3-none-any.whl
- Upload date:
- Size: 18.8 kB
- Tags: Python 2, Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.1 CPython/3.9.14
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 33fdd01ec046af06c43093a8fb42e756c47503244fcb8edfe4d19bb15e13bfdc |
|
MD5 | f0d78479556e53c37962f3a25cfc0359 |
|
BLAKE2b-256 | a22cf63a484887ad8e17c921cb5a4bf395c16f282cb2b8b066c1e024fe08bbf4 |