INT8 BERT base uncased finetuned STS-B
Post-training static quantization
PyTorch
This is an INT8 PyTorch model quantized with huggingface/optimum-intel through the usage of Intel® Neural Compressor.
The original fp32 model comes from the fine-tuned model textattack/bert-base-uncased-STS-B.
Test result
| INT8 | FP32 | |
|---|---|---|
| Accuracy (eval-f1) | 0.8755 | 0.8805 |
| Model size (MB) | 118 | 438 |
Load with optimum:
from optimum.intel import INCModelForSequenceClassification
model_id = "Intel/bert-base-uncased-STS-B-int8"
int8_model = INCModelForSequenceClassification.from_pretrained(model_id)