Open-source Universal-NER-UniNER-7B Model: Reduce Energy Consumption and Memory, Conduct Accurate Named Entity Recognition

Universal NER UniNER 7B All Bnb 4bit Smashed

Developed by PrunaAI

PrunaAI's compressed version of the UniNER-7B-all model, significantly reducing memory usage and energy consumption through quantization techniques while maintaining good named entity recognition capabilities.

Large Language Model

Transformers

#Named Entity Recognition #4bit Quantization #Low Memory Consumption

Downloads 22

Release Time : 4/12/2024

Model Overview

A 4-bit quantized compressed version based on Universal-NER/UniNER-7B-all, optimized for efficient named entity recognition tasks, suitable for deployment in resource-constrained environments.

Model Features

Efficient Compression

Utilizes llm-int8 and 4-bit quantization techniques to significantly reduce model memory usage.

Eco-friendly and Energy-saving

The optimized model reduces CO2 emissions and energy consumption during inference.

Plug-and-Play

Maintains the same interface as the original model, allowing deployment without modifying existing code.

Model Capabilities

Named Entity Recognition

Text Understanding

Multilingual Processing

Use Cases

Information Extraction

Automated Document Processing

Automatically extracts key entity information from legal documents or medical records.

Improves document processing efficiency and reduces manual labeling costs.

Content Analysis

Social Media Monitoring

Real-time identification of named entities (people, organizations, locations, etc.) in social media.

Supports brand monitoring and sentiment analysis.

🚀 PrunaAI: Simply make AI models cheaper, smaller, faster, and greener!

PrunaAI aims to optimize AI models, making them more cost - effective, compact, high - speed, and environmentally friendly. It provides compression solutions for various AI models and offers a series of metrics to evaluate the efficiency of these models.

🚀 Quick Start

👍 Give a thumbs up if you like this model!
📧 Contact us and tell us which model to compress next here.
📈 Request access to easily compress your own AI models here.
📖 Read the documentations to know more here
💬 Join Pruna AI community on Discord here to share feedback/suggestions or get help.

✨ Features

Metrics

memory_disk
memory_inference
inference_latency
inference_throughput
inference_CO2_emissions
inference_energy_consumption

📦 Installation

You can run the smashed model with these steps:

Check requirements from the original repo Universal-NER/UniNER-7B-all installed. In particular, check python, cuda, and transformers versions.
Make sure that you have installed quantization related packages.
```
pip install transformers accelerate bitsandbytes>0.37.0
```

Load & run the model.

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("PrunaAI/Universal-NER-UniNER-7B-all-bnb-4bit-smashed",
                                             trust_remote_code=True, device_map='auto')
tokenizer = AutoTokenizer.from_pretrained("Universal-NER/UniNER-7B-all")

input_ids = tokenizer("What is the color of prunes?,", return_tensors='pt').to(model.device)["input_ids"]

outputs = model.generate(input_ids, max_new_tokens=216)
tokenizer.decode(outputs[0])

📚 Documentation

Results

image info

Frequently Asked Questions

How does the compression work? The model is compressed with llm - int8.
How does the model quality change? The quality of the model output might vary compared to the base model.
How is the model efficiency evaluated? These results were obtained on NVIDIA A100 - PCIE - 40GB with configuration described in model/smash_config.json and are obtained after a hardware warmup. The smashed model is directly compared to the original base model. Efficiency results may vary in other settings (e.g. other hardware, image size, batch size, ...). We recommend to directly run them in the use - case conditions to know if the smashed model can benefit you.
What is the model format? We use safetensors.
What calibration data has been used? If needed by the compression method, we used WikiText as the calibration data.
What is the naming convention for Pruna Huggingface models? We take the original model name and append "turbo", "tiny", or "green" if the smashed model has a measured inference speed, inference memory, or inference energy consumption which is less than 90% of the original base model.
How to compress my own models? You can request premium access to more compression methods and tech support for your specific use - cases here.
What are "first" metrics? Results mentioning "first" are obtained after the first run of the model. The first run might take more memory or be slower than the subsequent runs due cuda overheads.
What are "Sync" and "Async" metrics? "Sync" metrics are obtained by syncing all GPU processes and stop measurement when all of them are executed. "Async" metrics are obtained without syncing all GPU processes and stop when the model output can be used by the CPU. We provide both metrics since both could be relevant depending on the use - case. We recommend to test the efficiency gains directly in your use - cases.

Configurations

The configuration info are in smash_config.json.

📄 License

The license of the smashed model follows the license of the original model. Please check the license of the original model Universal - NER/UniNER - 7B - all before using this model which provided the base model. The license of the pruna - engine is [here](https://pypi.org/project/pruna - engine/) on Pypi.

💡 More Compression Requests

📧 Contact us and tell us which model to compress next here.
📈 Request access to easily compress your own AI models [here](https://z0halsaff74.typeform.com/pruna - access?typeform - source=www.pruna.ai).

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご