Jina Embeddings V2 Base Code: An Open-Source Code Text Embedding Model - Freely Extract Code Features and Calculate Similarity

Jina Embeddings V2 Base Code GGUF

Developed by second-state

Jina Embeddings V2 Base Code is a transformer-based English sentence embedding model, specializing in feature extraction and similarity computation for code-related texts.

Text Embedding EnglishOpen Source License:Apache-2.0 #Code Embedding #High-Dimensional Vectors #Long Text Processing

Downloads 315

Release Time : 12/10/2024

Model Overview

This model is primarily used to generate high-quality sentence embeddings, particularly for processing code-related texts, supporting tasks like sentence similarity computation and feature extraction.

Model Features

Long Context Support

Supports long-context processing up to 8192 tokens.

High-Quality Embeddings

Generates high-quality 768-dimensional sentence embeddings.

Code Optimization

Specially optimized for code-related texts.

MTEB Benchmark Performance

Demonstrates excellent performance on the MTEB benchmark.

Model Capabilities

Sentence embedding generation

Text feature extraction

Sentence similarity computation

Code text processing

Use Cases

Code Search and Retrieval

Code Snippet Similarity Search

Find semantically similar code snippets in a codebase.

Improves code reuse and development efficiency.

Document Processing

Technical Document Clustering

Automatically classify and cluster technical documents.

Enhances document organization and retrieval efficiency.

🚀 jina-embeddings-v2-base-code-GGUF

This project provides quantized GGUF models based on the jina-embeddings-v2-base-code model, enabling efficient inference with LlamaEdge.

🚀 Quick Start

Original Model

The original model can be found at jinaai/jina-embeddings-v2-base-code.

Run with LlamaEdge

LlamaEdge version: v0.14.17
Prompt template:
- Prompt type: embedding
Context size: 8192
Embedding dim: 768

To run as a LlamaEdge service, use the following command:

wasmedge --dir .:. --nn-preload default:GGML:AUTO:jina-embeddings-v2-base-code-f16.gguf \
  llama-api-server.wasm \
  --prompt-template embedding \
  --ctx-size 8192 \
  --model-name jina-embeddings-v2-base-code

📦 Model Information

Property	Details
Base Model	jinaai/jina-embeddings-v2-base-code
License	apache-2.0
Model Creator	jinaai
Quantized By	Second State Inc.
Language	en
Inference	false
Tags	sentence-transformers, feature-extraction, sentence-similarity, mteb, transformers, transformers.js

📚 Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
jina-embeddings-v2-base-code-Q2_K.gguf	Q2_K	2	82.7 MB	smallest, significant quality loss - not recommended for most purposes
jina-embeddings-v2-base-code-Q3_K_L.gguf	Q3_K_L	3	101 MB	small, substantial quality loss
jina-embeddings-v2-base-code-Q3_K_M.gguf	Q3_K_M	3	95.6 MB	very small, high quality loss
jina-embeddings-v2-base-code-Q3_K_S.gguf	Q3_K_S	3	89.8 MB	very small, high quality loss
jina-embeddings-v2-base-code-Q4_0.gguf	Q4_0	4	105 MB	legacy; small, very high quality loss - prefer using Q3_K_M
jina-embeddings-v2-base-code-Q4_K_M.gguf	Q4_K_M	4	109 MB	medium, balanced quality - recommended
jina-embeddings-v2-base-code-Q4_K_S.gguf	Q4_K_S	4	105 MB	small, greater quality loss
jina-embeddings-v2-base-code-Q5_0.gguf	Q5_0	5	119 MB	legacy; medium, balanced quality - prefer using Q4_K_M
jina-embeddings-v2-base-code-Q5_K_M.gguf	Q5_K_M	5	121 MB	large, very low quality loss - recommended
jina-embeddings-v2-base-code-Q5_K_S.gguf	Q5_K_S	5	119 MB	large, low quality loss - recommended
jina-embeddings-v2-base-code-Q6_K.gguf	Q6_K	6	134 MB	very large, extremely low quality loss
jina-embeddings-v2-base-code-Q8_0.gguf	Q8_0	8	173 MB	very large, extremely low quality loss - not recommended
jina-embeddings-v2-base-code-f16.gguf	f16	16	323 MB	very large, extremely low quality loss - not recommended

Quantized with llama.cpp b4273

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