cuuupid/e5-mistral-7b-instruct 🖼️🔢✓ → 🖼️

▶️ 245 runs 📅 Feb 2024 ⚙️ Cog 0.9.4 📄 Paper ⚖️ License
semantic-search text-embedding

Performance

16.4sTypical run time
~260sCold start (first call)
245Total runs

About

Finetuned E5 embeddings for instruct based on Mistral.

Example Output

Output

Example output

Performance Metrics

16.40s Prediction Time
260.18s Total Time
All Input Parameters 
{
  "texts": "https://replicate.delivery/pbxt/KLrAjasFNPYHr7SKynYHT1bbC63uUs6TMGcCcvxgmFAh9fLJ/embedding-text.json",
  "batch_size": 1,
  "normalize_embeddings": true
}
Input Parameters
texts Type: string
Provide your input texts in a JSONL file, where each like is a JSON stringified string to embed.
batch_size Type: integerDefault: 32
The batch size to use for inference.
normalize_embeddings Type: booleanDefault: true
Whether to normalize the embeddings.
Output Schema

Output

Type: string • Format: uri

Example Execution Logs 
Max tokens: 9
Index: 0
Max tokens: 6
Index: 1
Max tokens: 5
Index: 2
Max tokens: 7
Index: 3
Max tokens: 11490
Index: 4
Version Details
Version ID
9452d5fc5d9b6ff5693ac89966fe19d755006a3a0367036f8f1244a5599ebdd1
Version Created
February 3, 2024
Run on Replicate →