---
title: albert-base-v2-STS-B
canonical_url: "https://www.modelscope.cn/models/textattack/albert-base-v2-STS-B"
md_url: "https://www.modelscope.cn/models/textattack/albert-base-v2-STS-B.md"
repository: textattack/albert-base-v2-STS-B
last_updated: 2025-09-28
pipeline_tag: text-classification
tasks:
  - text-classification
model_type:
  - albert
architectures:
  - AlbertForSequenceClassification
library_name:
  - pytorch
frameworks:
  - pytorch
downloads: 85
stars: 0
---

# albert-base-v2-STS-B

> albert-base-v2-STS-B - textattack 在 ModelScope 开源的模型。TextAttack Model Card This albert-base-v2 model was fine-tuned for sequence classification using TextAttack and the glue dataset loaded using the nlp library. The model was fine-tuned for 5 epochs with a…

- **Repository**: textattack/albert-base-v2-STS-B
- **Tasks**: text-classification
- **Downloads**: 85
- **Stars**: 0
- **Last updated**: 2025-09-28

Source: https://www.modelscope.cn/models/textattack/albert-base-v2-STS-B

---

## TextAttack Model Card
This `albert-base-v2` model was fine-tuned for sequence classification using TextAttack 
and the glue dataset loaded using the `nlp` library. The model was fine-tuned 
for 5 epochs with a batch size of 32, a learning 
rate of 3e-05, and a maximum sequence length of 128. 
Since this was a regression task, the model was trained with a mean squared error loss function. 
The best score the model achieved on this task was 0.9064220351504577, as measured by the 
eval set pearson correlation, found after 3 epochs.

For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack).