biubiubiu111/coral-surge-v2

Embedding Service

Sentence embeddings via all-MiniLM-L6-v2. Optimized for low-latency inference on CPU.

Features

• —Built-in tokenizer with padding and truncation
• —Input validation and sanitization
• —Model quantization support (INT8/FP16)

Usage

docker run -p 7860:7860 -e MODEL_NAME=distilbert-base IMAGE_NAME