|
|
--- |
|
|
language: |
|
|
- zh |
|
|
- en |
|
|
library_name: transformers |
|
|
pipeline_tag: text2text-generation |
|
|
--- |
|
|
|
|
|
抽取情感四元组、三元组、二元组等 |
|
|
```python |
|
|
import torch |
|
|
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("yuyijiong/mt0-xl-bf16-sentiment-quadruple") |
|
|
model = AutoModelForSeq2SeqLM.from_pretrained("yuyijiong/mt0-xl-bf16-sentiment-quadruple", device_map="auto", torch_dtype=torch.bfloat16) |
|
|
|
|
|
text = '情感四元组(对象,观点,方面,极性)抽取任务 (补全null): 【已经开了4袋,普遍出现米沾在包装上了,看起来放了很久的样子】' |
|
|
input_ids = tokenizer(text,return_tensors="pt", padding=True)['input_ids'].cuda(0) |
|
|
with torch.no_grad(): |
|
|
with torch.autocast('cuda'): |
|
|
output = model.generate(input_ids=input_ids) |
|
|
output_str = tokenizer.batch_decode(output, skip_special_tokens=True) |
|
|
print(output_str) |
|
|
|
|
|
``` |
