Zeroshot (Python)
Image classification for the masses
Installation
Install via pip: pip install zeroshot
For GPU support, pip install zeroshot[torch]
N.B. In theory ONNX supports GPU, but the restrictions on CUDA version are iffy at best, and so for easiest results just use PyTorch. If you're brave, instead pip install onnxruntime-gpu.
Usage
First, go to app.usezeroshot.com and create a classifier. Check out the video on the landing page for an example.
Then, in Python (image should be an RGB numpy array with channels last):
import zeroshot
# Create the classifier and preprocessing function.
classifier = zeroshot.Classifier("model-uuid-goes-here")
preprocess_fn = zeroshot.create_preprocess_fn()
# Run the model!
prediction = classifier.predict(preprocess_fn(image))
print(f"The image is class {prediction}")
You can also download the classifier and save it somewhere locally so you don't need to hit the server each time. Hit "download model" in the web-app and save the json file somewhere. You can then instead do:
classifier = zeroshot.Classifier("/home/user/path/to/model.json")
Additional Tips
- To use a GPU, install the torch backend with
pip install zeroshot[torch] - If you are hitting issues with torch trying to run on CPU, try disabling XFormers by setting XFORMERS_DISABLED=1 in your ENV varaibles.
Read the docs
See the docs folder for some details on how things work under the hood.
Get help
If you need help or just want to chat, join the Moonshine Labs Slack server and come hang out in the #zeroshot channel.
Metadata
Release files for zeroshot 0.1.11
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 |
|---|---|---|---|---|
| zeroshot-0.1.11.tar.gz | 24.4 kB | Without provenance | Details for zeroshot-0.1.11.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| zeroshot-0.1.11-py3-none-any.whl | Python 3 | none | any | Details for zeroshot-0.1.11-py3-none-any.whl |
Total release size: 50.8 kB
Release files / zeroshot-0.1.11.tar.gz
| Download URL | zeroshot-0.1.11.tar.gz |
|---|---|
| Size | 24.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
7e012ac58ccb1979560f294846d86c475d05f9bfe660e948e15294296b37522b
|
|
BLAKE2b-256 checksum How to use checksums |
d3534ca17cfe405bce9e9d33b657c7d886d3c05cd942265e145af61026bb10f6
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.6.1 CPython/3.9.10 Linux/6.2.0-37-generic
|
Release files / zeroshot-0.1.11-py3-none-any.whl
| Download URL | zeroshot-0.1.11-py3-none-any.whl |
|---|---|
| Size | 26.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
208452598d4426b64f78f1c9c3d4caf0b314bc10a55657c38384c3f77d946b9f
|
|
BLAKE2b-256 checksum How to use checksums |
01b8118650fa88ce84de4897229bce0026600657d497807fe4b066ea4727c1ef
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.6.1 CPython/3.9.10 Linux/6.2.0-37-generic
|
Security for zeroshot 0.1.11
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.