|
|
--- |
|
|
license: apache-2.0 |
|
|
datasets: |
|
|
- CarperAI/openai_summarize_tldr |
|
|
metrics: |
|
|
-rouge1: 0.3156 |
|
|
--- |
|
|
|
|
|
### Example usage |
|
|
|
|
|
<p> |
|
|
|
|
|
</p> |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("gpt2") |
|
|
model = AutoModelForCausalLM.from_pretrained("KookyGhost/GPT2-small-summarization") |
|
|
|
|
|
prompt = "Summarize this: Reddit user shares a long story about learning to code with free online resources and eventually landing their first developer job." |
|
|
|
|
|
inputs = tokenizer(prompt, return_tensors="pt", truncation=True) |
|
|
outputs = model.generate( |
|
|
**inputs, |
|
|
max_new_tokens=60, |
|
|
do_sample=True, |
|
|
top_k=50, |
|
|
top_p=0.95, |
|
|
temperature=0.7 |
|
|
) |
|
|
|
|
|
summary = tokenizer.decode(outputs[0], skip_special_tokens=True) |
|
|
print(summary) |
|
|
|
