Update README.md
Browse files
README.md
CHANGED
|
@@ -72,6 +72,22 @@ We compared the model against `all-minilm-l6-v2`/`all-mpnet-base-v2` from sbert
|
|
| 72 |
|jina-embedding-b-en-v1|0.735|0.792|0.752|0.851|0.801|0.89|0.546 |0.871|0.586 |
|
| 73 |
|jina-embedding-l-en-v1|0.739|**0.844**|0.778|**0.863**|0.821|0.896|0.566 |**0.882**|0.608 |
|
| 74 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
## Usage
|
| 76 |
|
| 77 |
Use with Jina AI Finetuner
|
|
|
|
| 72 |
|jina-embedding-b-en-v1|0.735|0.792|0.752|0.851|0.801|0.89|0.546 |0.871|0.586 |
|
| 73 |
|jina-embedding-l-en-v1|0.739|**0.844**|0.778|**0.863**|0.821|0.896|0.566 |**0.882**|0.608 |
|
| 74 |
|
| 75 |
+
## Inference Speed
|
| 76 |
+
|
| 77 |
+
We encoded a single sentence `What is the current weather like today?` 10k times on:
|
| 78 |
+
|
| 79 |
+
1. cpu: MacBook Pro 2020, 2 GHz Quad-Core Intel Core i5
|
| 80 |
+
2. gpu: 1 Nvidia 3090
|
| 81 |
+
|
| 82 |
+
And recorded time spent to demonstrate the embedding speed:
|
| 83 |
+
|
| 84 |
+
|Name|param |dimension| time@cpu | time@gpu |
|
| 85 |
+
|------------------------------|-----|------|-----|-----|
|
| 86 |
+
|jina-embedding-t-en-v1|14m |312| 5.78s | |
|
| 87 |
+
|all-minilm-l6-v2|23m |384| 11.95s | |
|
| 88 |
+
|jina-embedding-s-en-v1|35m |512| 17.25s | |
|
| 89 |
+
|
| 90 |
+
|
| 91 |
## Usage
|
| 92 |
|
| 93 |
Use with Jina AI Finetuner
|