MultiModalCrossAttn
The open source implementation of the cross attention mechanism from the paper: "JOINTLY TRAINING LARGE AUTOREGRESSIVE MULTIMODAL MODELS"
Appreciation
- Lucidrains
- Agorians
Install
pip install cross-attn
Usage
import torch
from cross_attn.main import MultiModalCrossAttention
# Test the MultiModalCrossAttention module
dim = 512 # For example
num_heads = 8
cross_attn = MultiModalCrossAttention(dim, num_heads)
Hllm_sample = torch.randn(32, dim, dim) # Batch size = 32, Sequence length = 10
Himg_sample = torch.randn(32, dim, dim)
output = cross_attn(Hllm_sample, Himg_sample)
print(output)
print(output.shape) # Expected: [32, 10, 512]
License
MIT
Citations
@misc{2309.15564,
Author = {Emanuele Aiello and Lili Yu and Yixin Nie and Armen Aghajanyan and Barlas Oguz},
Title = {Jointly Training Large Autoregressive Multimodal Models},
Year = {2023},
Eprint = {arXiv:2309.15564},
}
1 maintainer
Metadata
1 maintainer
Release files for cross-attn 0.0.5
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 | Actions |
|---|---|---|---|---|
| cross_attn-0.0.5.tar.gz | 8.9 kB | Without provenance | Details for cross_attn-0.0.5.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| cross_attn-0.0.5-py3-none-any.whl | Python 3 | none | any | Details for cross_attn-0.0.5-py3-none-any.whl |
Total release size: 21.1 kB
Release files / cross_attn-0.0.5.tar.gz
| Download URL | cross_attn-0.0.5.tar.gz |
|---|---|
| Size | 8.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
588b54dc7ea9e1515286977956ceadf0b81711c2ab2b0e8acb797c01d331ceba
|
|
BLAKE2b-256 checksum How to use checksums |
264511e7662ce13482e0d36438abbaea96e841af173d2c30a07a9072aaf405db
|
| 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 / cross_attn-0.0.5-py3-none-any.whl
| Download URL | cross_attn-0.0.5-py3-none-any.whl |
|---|---|
| Size | 12.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
ec2023b77e2be8ee15f40fe521f08d6b96700e4c443c751c33b3df9bf031baa8
|
|
BLAKE2b-256 checksum How to use checksums |
7b07a59c7c5fb4a38df1d13d98d4178f37ee8c650c97fbb7b92d72c8c1bd85b8
|
| 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
|
Security for cross-attn 0.0.5
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.
