gte-Qwen2-7B-instruct | PromptLayer Models
Webinar · Jul 20
Fine-tuning open-source models: is it time to move off Frontier Lab models?
gte-Qwen2-7B-instruct
Alibaba-NLP
State-of-the-art multilingual embedding model ranking #1 on MTEB benchmark, built on Qwen2-7B with 3584-dim embeddings and 32k context window
| Property | Value |
|---|---|
| Parameter Count | 7 billion |
| Embedding Dimension | 3584 |
| Max Context Length | 32,000 tokens |
| License | Apache 2.0 |
| Paper | Research Paper |
What is gte-Qwen2-7B-instruct?
gte-Qwen2-7B-instruct represents the latest advancement in the General Text Embedding (GTE) model family, achieving benchmark-leading performance in both English and Chinese evaluations on the MTEB leaderboard. Built upon Qwen's powerful 7B parameter architecture, this model introduces significant improvements over its predecessor through enhanced bidirectional attention mechanisms and comprehensive multilingual training.
Implementation Details
The model leverages sophisticated architectural choices including bidirectional attention for improved context understanding and instruction tuning specifically optimized for query processing. With its 3584-dimensional embedding space and extensive 32k token context window, it offers robust capability for handling complex text relationships.
- State-of-the-art performance on MTEB (70.24) and C-MTEB (72.05) benchmarks
- Comprehensive multilingual support with strong performance across English, Chinese, French, and Polish
- Advanced instruction tuning focused on query optimization
- Built on the improved Qwen2 architecture for enhanced representation learning
Core Capabilities
- Superior multilingual text embedding generation
- Long-context understanding with 32k token support
- Efficient semantic search and retrieval
- Cross-lingual text similarity analysis
- Document classification and clustering
Frequently Asked Questions
Q: What makes this model unique?
The model's distinctive feature is its combination of Qwen2's advanced architecture with specialized instruction tuning, resulting in state-of-the-art performance across multiple languages while maintaining efficient processing through query-side optimization.
Q: What are the recommended use cases?
The model excels in semantic search, document retrieval, text classification, and cross-lingual applications. It's particularly effective for enterprise-scale applications requiring robust multilingual understanding and high-precision text matching.