|
|
--- |
|
|
language: en |
|
|
license: apache-2.0 |
|
|
tags: text-classfication |
|
|
datasets: |
|
|
- sst2 |
|
|
--- |
|
|
|
|
|
INT8 DistilBERT base uncased finetuned SST-2 (Post-training static quantization) |
|
|
=== |
|
|
This is an INT8 PyTorch model quantized by [intel/nlp-toolkit](https://github.com/intel/nlp-toolkit) using provider: [Intel® Neural Compressor](https://github.com/intel/neural-compressor). The original fp32 model comes from the fine-tuned model [distilbert-base-uncased-finetuned-sst-2-english](https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english) |
|
|
|
|
|
Test result below comes from [AWS](https://aws.amazon.com/) c6i.xlarge (intel ice lake: 4 vCPUs, 8g Memory) instance. |
|
|
|
|
|
| |fp32|int8| |
|
|
|---|:---:|:---:| |
|
|
| **Accuracy** |0.9106|0.9037| |
|
|
| **Throughput (samples/sec)** |？|？| |
|
|
| **Model size (MB)** |255|66| |
|
|
|
|
|
|
|
|
Load with optimum: |
|
|
```python |
|
|
from nlp_toolkit import OptimizedModel |
|
|
int8_model = OptimizedModel.from_pretrained( |
|
|
'intel/distilbert-base-uncased-finetuned-sst-2-english-int8-static', |
|
|
) |
|
|
``` |
|
|
Notes: |
|
|
- The INT8 model has better performance than the FP32 model when the CPU is fully loaded. Otherwise, there will be the illusion that INT8 is inferior to FP32. |
|
|
|
