|
--- |
|
license: bigscience-bloom-rail-1.0 |
|
language: |
|
- zh |
|
pipeline_tag: text-generation |
|
widget: |
|
- text: "中国的首都是" |
|
|
|
--- |
|
|
|
This model is based on [bigscience/bloom-1b7](https://huggingface.co/bigscience/bloom-1b7). |
|
|
|
We pruned its vocabulary from 250880 to 46145 with Chinese corpus to reduce GPU memory usage. So the total parameter is 1.4b now. |
|
|
|
# How to use |
|
```python |
|
from transformers import BloomTokenizerFast, BloomForCausalLM |
|
|
|
tokenizer = BloomTokenizerFast.from_pretrained('Langboat/bloom-1b4-zh') |
|
model = BloomForCausalLM.from_pretrained('Langboat/bloom-1b4-zh') |
|
|
|
print(tokenizer.batch_decode(model.generate(tokenizer.encode('中国的首都是', return_tensors='pt')))) |
|
``` |
