Encoding Speed on Encoding Speed Benchmark 512 tokens
3,113Encoding Speed (Docs/s)granite-embedding-107m-multilingual
Evaluation results
| Method | Links | ||
|---|---|---|---|
| granite-embedding-107m-multilingualParameters (M)=107, Embedding Size=384, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 3,113 | 122.9 | |
| multilingual-e5-smallParameters (M)=96, Embedding Size=384, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,604 | 118.9 | |
| granite-embedding-97m-multilingual-r2Parameters (M)=97, Embedding Size=384, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,534 | 100 | |
| F2LLM-v2-80MParameters (M)=80, Embedding Size=320, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,190 | 86.4 | |
| snowflake-arctic-embed-m-v2.0Parameters (M)=305, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,190 | 119.8 | |
| granite-embedding-278m-multilingualParameters (M)=278, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,164 | 118.4 | |
| multilingual-e5-baseParameters (M)=278, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,025 | 110.8 | |
| gte-multilingual-baseParameters (M)=305, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 2,018 | 110.4 | |
| harrier-oss-v1-270mParameters (M)=270, Embedding Size=640, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 1,938 | 106 | |
| granite-embedding-311m-multilingual-r2Parameters (M)=311, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 1,828 | 100 | |
| embeddinggemma-300mParameters (M)=300, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 1,349 | 73.8 | |
| jina-embeddings-v5-text-nanoParameters (M)=239, Embedding Size=768, Hardware=Nvidia H100 GPU, Batch Size=512, Max Tokens=512, Software Version=HuggingFace transformers 5.8.02026.05 | 307 | 16.8 |