Model
ZeroShotIntentModel trained by fine tuning bert-base-uncased on the MNLI (Multi-Genre Natural Language Inference) dataset, which achieves an accuracy of 84.9% and 84.8% on the matched and mismatched dev sets, respectively.
Use the NGC CLI to download:
Copied!
1 Version
1.4.1Selected10/07/2021 11:54 AM UTC388.86 MB Copied!
1.4.1Selected
10/07/2021 11:54 AM UTC388.86 MB
Copied!
Accuracy
| Key | Value |
|---|---|
| MNLI dev set (matched) | 84.9% |
| MNLI dev set (mismatched) | 84.8% |
Model
| Key | Value |
|---|---|
| Architecture | BERT base uncased |
| Inputs | list of strings, list of strings |
| Outputs | list of dictionaries |