Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/backprop-ai/backprop
Backprop makes it simple to use, finetune, and deploy state-of-the-art ML models.
https://github.com/backprop-ai/backprop
bert fine-tuning image-classification language-model multilingual-models natural-language-processing nlp question-answering text-classification transfer-learning transformers
Last synced: 3 months ago
JSON representation
Backprop makes it simple to use, finetune, and deploy state-of-the-art ML models.
- Host: GitHub
- URL: https://github.com/backprop-ai/backprop
- Owner: backprop-ai
- License: other
- Created: 2020-10-30T15:25:14.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2021-05-03T09:15:25.000Z (over 3 years ago)
- Last Synced: 2024-10-13T15:48:40.217Z (4 months ago)
- Topics: bert, fine-tuning, image-classification, language-model, multilingual-models, natural-language-processing, nlp, question-answering, text-classification, transfer-learning, transformers
- Language: Python
- Homepage: https://backprop.co
- Size: 5.46 MB
- Stars: 243
- Watchers: 16
- Forks: 12
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-pytorch-list-CNVersion - Backprop - of-the-art ML models. (Pytorch & related libraries｜Pytorch & 相关库 / NLP & Speech Processing｜自然语言处理 & 语音处理:)
- awesome-production-machine-learning - Backprop - ai/backprop.svg?style=social) - Backprop makes it simple to use, finetune, and deploy state-of-the-art ML models. (Deployment and Serving)
- Awesome-pytorch-list - Backprop - of-the-art ML models. (Pytorch & related libraries / NLP & Speech Processing:)
README
Backprop makes it simple to use, finetune, and deploy state-of-the-art ML models.
Solve a variety of tasks with pre-trained models or finetune them in one line for your own tasks.
Out of the box tasks you can solve with Backprop:
- Conversational question answering in English
- Text Classification in 100+ languages
- Image Classification
- Text Vectorisation in 50+ languages
- Image Vectorisation
- Summarisation in English
- Emotion detection in English
- Text Generation
For more specific use cases, you can adapt a task with little data and a single line of code via finetuning.
| ⚡ [Getting started](#getting-started) | Installation, few minute introduction |
| :---------------------------------------------------- | :-------------------------------------------------------- |
| 💡 [Examples](#examples) | Finetuning and usage examples |
| 📙 [Docs](https://backprop.readthedocs.io/en/latest/) | In-depth documentation about task inference and finetuning |
| ⚙️ [Models](https://backprop.co/hub) | Overview of available models |
## Getting started
### Installation
Install Backprop via PyPi:
```bash
pip install backprop
```
### Basic task inference
Tasks act as interfaces that let you easily use a variety of supported models.
```python
import backprop
context = "Take a look at the examples folder to see use cases!"
qa = backprop.QA()
# Start building!
answer = qa("Where can I see what to build?", context)
print(answer)
# Prints
"the examples folder"
```
You can run all tasks and models on your own machine, or in production with our inference [API](https://backprop.co), simply by specifying your `api_key`.
See how to use [all available tasks](https://backprop.readthedocs.io/en/latest/Tasks.html).
### Basic finetuning and uploading
Each task implements finetuning that lets you adapt a model for your specific use case in a single line of code.
A finetuned model is easy to upload to production, letting you focus on building great applications.
```python
import backprop
tg = backprop.TextGeneration("t5-small")
# Any text works as training data
inp = ["I really liked the service I received!", "Meh, it was not impressive."]
out = ["positive", "negative"]
# Finetune with a single line of code
tg.finetune({"input_text": inp, "output_text": out})
# Use your trained model
prediction = tg("I enjoyed it!")
print(prediction)
# Prints
"positive"
# Upload to Backprop for production ready inference
# Describe your model
name = "t5-sentiment"
description = "Predicts positive and negative sentiment"
tg.upload(name=name, description=description, api_key="abc")
```
See [finetuning for other tasks](https://backprop.readthedocs.io/en/latest/Finetuning.html).
## Why Backprop?
1. No experience needed
- Entrance to practical AI should be simple
- Get state-of-the-art performance in your task without being an expert
2. Data is a bottleneck
- Solve real world tasks without any data
- With transfer learning, even a small amount of data can adapt a task to your niche requirements
3. There are an overwhelming amount of models
- We offer a curated selection of the best open-source models and make them simple to use
- A few general models can accomplish more with less optimisation
4. Deploying models cost effectively is hard work
- If our models suit your use case, no deployment is needed: just call our API
- Adapt and deploy your own model with just a few lines of code
- Our API scales, is always available, and you only pay for usage
## Examples
- Solve any text based task with Finetuning ([Github](https://github.com/backprop-ai/backprop/blob/main/examples/Finetuning_GettingStarted.ipynb), [Colab](https://colab.research.google.com/github/backprop-ai/backprop/blob/main/examples/Finetuning_GettingStarted.ipynb))
- Search for images using text ([Github](https://github.com/backprop-ai/backprop/blob/main/examples/ImageVectorisation.ipynb))
- Finding answers from text ([Github](https://github.com/backprop-ai/backprop/blob/main/examples/Q%26A.ipynb))
- [More finetuning and task examples](https://github.com/backprop-ai/backprop/tree/main/examples)
## Documentation
Check out our [docs](https://backprop.readthedocs.io/en/latest/) for in-depth task inference and finetuning.
## Model Hub
Curated list of [state-of-the-art models](https://backprop.co/hub).
## Demos
Zero-shot image classification with [CLIP](https://clip.backprop.co).
## Credits
Backprop relies on many great libraries to work, most notably:
* [PyTorch](https://github.com/pytorch/pytorch)
* [PyTorch Lightning](https://github.com/PyTorchLightning/pytorch-lightning)
* [Transformers](https://github.com/huggingface/transformers)
* [Sentence Transformers](https://github.com/UKPLab/sentence-transformers)
* [EfficientNet PyTorch](https://github.com/lukemelas/EfficientNet-PyTorch)
* [CLIP](https://github.com/openai/CLIP)
## Feedback
Found a bug or have ideas for new tasks and models? Open an [issue](https://github.com/backprop-ai/backprop/issues).