| language: multilingual | |
| library_name: sentence-transformers | |
| pipeline_tag: sentence-similarity | |
| tags: | |
| - sentence-transformers | |
| - multilingual | |
| - nutrition | |
| - retrieval | |
| base_model: intfloat/multilingual-e5-base | |
| # Multilingual E5 Nutrition Fine-tuned | |
| This model is a fine-tuned version of `intfloat/multilingual-e5-base` specifically for nutrition domain retrieval. | |
| ## Model Details | |
| - Base Model: intfloat/multilingual-e5-base | |
| - Fine-tuned on: 5228 nutrition samples | |
| - Best Validation Loss: 0.1619 | |
| - Architecture: Identical to original multilingual-e5-base | |
| ## Usage | |
| Use exactly like the original multilingual-e5-base: | |
| |||python | |
| from sentence_transformers import SentenceTransformer | |
| model = SentenceTransformer('linhha2705/multilingual-e5-simple') | |
| # Encode queries and passages with prefixes | |
| query = "query: What is healthy food?" | |
| passage = "passage: Healthy food includes fruits and vegetables." | |
| query_embedding = model.encode(query) | |
| passage_embedding = model.encode(passage) | |
| ||| | |
| ## Performance | |
| Improved retrieval accuracy for nutrition domain while maintaining full compatibility with multilingual-e5-base. | |