Greedy Decoder in default SmolLM2-360M prevents num_return_sequences=2 and gives an error

In ch 1.3 [Transformers: What can they do](https://huggingface.co/learn/llm-course/chapter1/3#using-any-model-from-the-hub-in-a-pipeline), the first example under "Using any model from the Hub in a pipeline" gives an error because the model by default is using Greedy Decoder and hence cannot generate 2 sequences. Here is the current example code:

```
from transformers import pipeline

generator = pipeline("text-generation", model="HuggingFaceTB/SmolLM2-360M")
generator(
"In this course, we will teach you how to",
max_length=30,
num_return_sequences=2,
)
```

Suggestion: Include num_beams or do_sampling=True