Metadata-Version: 2.1
Name: jamba
Version: 0.0.2
Summary: jamba - Pytorch
Home-page: https://github.com/kyegomez/jamba
License: MIT
Keywords: artificial intelligence,deep learning,optimizers,Prompt Engineering
Author: Kye Gomez
Author-email: kye@apac.ai
Requires-Python: >=3.6,<4.0
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.6
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.9
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Dist: swarms
Requires-Dist: torch
Requires-Dist: zetascale
Project-URL: Documentation, https://github.com/kyegomez/jamba
Project-URL: Repository, https://github.com/kyegomez/jamba
Description-Content-Type: text/markdown

[![Multi-Modality](agorabanner.png)](https://discord.gg/qUtxnK2NMf)

# Jamba
PyTorch Implementation of Jamba: "Jamba: A Hybrid Transformer-Mamba Language Model"


## install
`$ pip install jamba`

## usage

```python
# Import the torch library, which provides tools for machine learning
import torch

# Import the Jamba model from the jamba.model module
from jamba.model import Jamba

# Create a tensor of random integers between 0 and 100, with shape (1, 100)
# This simulates a batch of tokens that we will pass through the model
x = torch.randint(0, 100, (1, 100))

# Initialize the Jamba model with the specified parameters
# dim: dimensionality of the input data
# depth: number of layers in the model
# num_tokens: number of unique tokens in the input data
# d_state: dimensionality of the hidden state in the model
# d_conv: dimensionality of the convolutional layers in the model
# heads: number of attention heads in the model
# num_experts: number of expert networks in the model
# num_experts_per_token: number of experts used for each token in the input data
model = Jamba(
    dim=512,
    depth=6,
    num_tokens=100,
    d_state=256,
    d_conv=128,
    heads=8,
    num_experts=8,
    num_experts_per_token=2,
)

# Perform a forward pass through the model with the input data
# This will return the model's predictions for each token in the input data
output = model(x)

# Print the model's predictions
print(output)

```

# License
MIT

