text-embeddings-inference
text-embeddings-inference copied to clipboard
Suggest supporting thenlper/gte-small-zh
System Info
Question: The reasoning results of TEI and sentence-transformers are inconsistent.
TEI version: 1.2 sentence-transformers: 2.3.1 embedding-model: thenlper/gte-small-zh
Information
- [X] Docker
- [ ] The CLI directly
Tasks
- [X] An officially supported command
- [ ] My own modifications
Reproduction
input text: 你好
TEI output: [[ 0.042778727, 0.016915003, 0.06579812, 0.04357415, -0.05951346, -0.028851831, ...]]
sentence-transformer output: [ 0.00476761, -0.04274084, -0.05193182, -0.01289688, -0.07046074, 0.01860032, ...]
Expected behavior
The outputs of TEI and sentence-transformers should be the same.