FRIDA GGUF

🚀 Model Card for FRIDA GGUF

FRIDA is a comprehensive fine - tuned general text embedding model. Inspired by the denoising architecture based on T5, it builds on the encoder part of the FRED - T5 model. The model continues the research on text embedding models, such as ruMTEB and [ru - en - RoSBERTa](https://huggingface.co/ai - forever/ru - en - RoSBERTa). It was pre - trained on a Russian - English dataset and fine - tuned for better performance on the target task.

• https://huggingface.co/evilfreelancer/FRIDA - GGUF
• https://ollama.com/evilfreelancer/FRIDA

For more model details, please refer to our technical report [TODO].

🚀 Quick Start

The model can be used with prefixes, and it's recommended to use CLS pooling. The choice of prefix and pooling depends on the task.

✨ Features

• Versatile Prefix Usage: Different prefixes are available for various tasks, such as retrieval, paraphrasing, categorization, and more.
• Fine - Tuning Capability: Can be fine - tuned with high - quality Russian and English datasets to better suit specific needs.

📦 Installation

Ollama

ollama pull evilfreelancer/FRIDA:f16

💻 Usage Examples

Basic Usage

We use the following basic rules to choose a prefix:

• "search_query: " and "search_document: " prefixes are for answer or relevant paragraph retrieval
• "paraphrase: " prefix is for symmetric paraphrasing related tasks (STS, paraphrase mining, deduplication)
• "categorize: " prefix is for asymmetric matching of document title and body (e.g. news, scientific papers, social posts)
• "categorize_sentiment: " prefix is for any tasks that rely on sentiment features (e.g. hate, toxic, emotion)
• "categorize_topic: " prefix is intended for tasks where you need to group texts by topic
• "categorize_entailment: " prefix is for textual entailment task (NLI)

Advanced Usage

Below are examples of texts encoding using the Transformers and SentenceTransformers libraries.

import json
import requests
import numpy as np

OLLAMA_HOST = "http://localhost:11434"
MODEL_NAME = "evilfreelancer/FRIDA:f16"


def get_embedding(text):
    payload = {
        "model": MODEL_NAME,
        "input": text
    }

    response = requests.post(
        f"{OLLAMA_HOST}/api/embed",
        data=json.dumps(payload, ensure_ascii=False),
        headers={"Content - Type": "application/x - www - form - urlencoded"}
    )
    response.raise_for_status()
    return np.array(response.json()["embeddings"][0])


def normalize(vectors):
    vectors = np.atleast_2d(vectors)
    norms = np.linalg.norm(vectors, axis=1, keepdims=True)
    norms[norms == 0] = 1.0
    return vectors / norms


def cosine_diag_similarity(a, b):
    return np.sum(a * b, axis=1)


inputs = [
    #
    "paraphrase: В Ярославской области разрешили работу бань, но без посетителей",
    "categorize_entailment: Женщину доставили в больницу за ее жизнь сейчас борются врачи.",
    "search_query: Сколько программистов нужно, чтобы вкрутить лампочку?",
    #
    "paraphrase: Ярославским баням разрешили работать без посетителей",
    "categorize_entailment: Женщину спасают врачи.",
    "search_document: Чтобы вкрутить лампочку нужно три программиста.",
]
size = int(len(inputs)/2)

embeddings = normalize(np.array([get_embedding(text) for text in inputs]))
sim_scores = cosine_diag_similarity(embeddings[:size], embeddings[size:])
print(sim_scores.tolist())

📚 Documentation

To better tailor the model to your needs, you can fine - tune it with relevant high - quality Russian and English datasets.

🔧 Technical Details

The model is based on the encoder part of FRED - T5 model and continues research on text embedding models. It has been pre - trained on a Russian - English dataset and fine - tuned for improved performance on the target task.

📄 License

This model is released under the MIT license.

Authors

• SaluteDevices AI for B2C RnD Team.
• Artem Snegirev: HF profile, Github;
• Anna Maksimova HF profile;
• Aleksandr Abramov: HF profile, Github, Kaggle Competitions Master
• Pavel Rykov: HF profile, Github - creator of GGUF version

Citation

@misc{TODO
}

Limitations

⚠️ Important Note

The model is designed to process texts in Russian, and the quality in English is unknown. The maximum input text length is limited to 512 tokens.