Commit
·
e02999f
1
Parent(s):
14a60c2
Update README.md
Browse files
README.md
CHANGED
|
@@ -26,11 +26,15 @@ model-index:
|
|
| 26 |
This model is a statically quantized version of [optimum/distilbert-base-uncased-finetuned-banking77](https://huggingface.co/optimum/distilbert-base-uncased-finetuned-banking77) on the `banking77` dataset.
|
| 27 |
It achieves the following results on the evaluation set:
|
| 28 |
|
|
|
|
|
|
|
| 29 |
- Vanilla model: 92.5%
|
| 30 |
- Quantized model: 92.24%
|
| 31 |
-
=> The quantized model achieves 99.72% accuracy of the fp32 model
|
| 32 |
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
| 34 |
Payload sequence length: 128
|
| 35 |
Instance type: AWS c6i.xlarge
|
| 36 |
|
|
|
|
| 26 |
This model is a statically quantized version of [optimum/distilbert-base-uncased-finetuned-banking77](https://huggingface.co/optimum/distilbert-base-uncased-finetuned-banking77) on the `banking77` dataset.
|
| 27 |
It achieves the following results on the evaluation set:
|
| 28 |
|
| 29 |
+
**Accuracy**
|
| 30 |
+
|
| 31 |
- Vanilla model: 92.5%
|
| 32 |
- Quantized model: 92.24%
|
|
|
|
| 33 |
|
| 34 |
+
> => The quantized model achieves 99.72% accuracy of the fp32 model
|
| 35 |
+
|
| 36 |
+
**Latency**
|
| 37 |
+
|
| 38 |
Payload sequence length: 128
|
| 39 |
Instance type: AWS c6i.xlarge
|
| 40 |
|