|
|
--- |
|
|
datasets: |
|
|
- upb-nlp/article_to_search_query |
|
|
language: |
|
|
- ro |
|
|
- en |
|
|
base_model: |
|
|
- OpenLLM-Ro/RoLlama2-7b-Base |
|
|
--- |
|
|
|
|
|
<div align="center"> |
|
|
<img src="images/logo.png" alt="Logo" width="240" height="240"> |
|
|
</div> |
|
|
|
|
|
## How to Get Started with the Model |
|
|
|
|
|
Use the code below to get started with the model. |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("upb-nlp/rollama2_7b_article_to_search_query") |
|
|
model = AutoModelForCausalLM.from_pretrained("upb-nlp/rollama2_7b_article_to_search_query") |
|
|
|
|
|
BASE_PROMPT = """You are a tool that turns news articles into realistic Google search queries someone might use to find the article. |
|
|
|
|
|
<article> |
|
|
{} |
|
|
</article> |
|
|
|
|
|
search query: """ |
|
|
|
|
|
INPUT_ARTICLE = "This is your article's title. This is your article's body." |
|
|
input_text = BASE_PROMPT.format(INPUT_ARTICLE) |
|
|
|
|
|
input_ids = tokenizer(input_text, truncation=True, max_length=1024, return_tensors="pt").to(torch.device('cuda')) |
|
|
outputs = model.generate(**input_ids, max_new_tokens=100) |
|
|
decoded_output = tokenizer.decode(outputs[0]) |
|
|
``` |
