You do the modelling. We do everything else.

We make it easy to launch machine learning models in the cloud and scale up as you grow – with serverless GPU’s, model versioning, shadow deployments and more.

pip install cerebrium
1
from cerebrium import deploy
2
import torch as nn
3

4
model = nn.Sequential(nn.Linear(n_input, n_hidden),
5
                      nn.ReLU,
6
                      nn.Linear(n_hidden, n_out),
7
                      nn.Sigmoid())
8
}
9
loss_function = nn.MSELoss()
optimizer = torch.optim.SGD(model.parameters(),lr=0.01)
10
11
some training...
12
13
torch.save(model.state_dict(), PATH)
14
model.load_state_dict(torch.load(PATH))
15
deploy(model, 'torch', MODEL_NAME, API_KEY)

Backed by

Ready to get  started?

Hobby

Free

No credit card required
folder_open
Unlimited models
edit_calendar
1 User
pending_actions
Serverless CPU/GPU
trending_up
1 hour free GPU
post_add
Automatic Versioning

Pro

Contact Us

plus compute down to the millisecond
folder_open
Unlimited models
edit_calendar
5 users
pending_actions
Serverless CPU/GPU
trending_up
1 hour free GPU
post_add
A/B Testing

Enterprise

Contact Us

billed annually
folder_open
Everything in Pro
edit_calendar
Unlimited Users
pending_actions
Scaled CPU/GPU Pricing
trending_up
SSO/SAML
post_add
Dedicated Manager