# \<unk\> token in the output instead curly braces

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Funk-token-in-the-output-instead-curly-braces%2F34653%3E%5C
**Category:** 🤗Tokenizers\
**Created:** [March 25, 2023, 6:49pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Funk-token-in-the-output-instead-curly-braces%2F34653 "2023-03-25T18:49:31Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![MihoZaki](https://avatars.discourse-cdn.com/v4/letter/m/da6949/32.png) [@MihoZaki](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FMihoZaki)%5C
**Post date:** [March 25, 2023, 6:49pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Funk-token-in-the-output-instead-curly-braces%2F34653%2F1 "2023-03-25T18:49:31Z")

</div>

Hello everyone! hope you all are well. i need some help because to be honest im super beginner and it got a bit confusing.  
I’m fine-tuning T5 for the task of generating mongodb queries as an ouput. i finished the training and when i loaded the model to test it, the output was missing Curly braces.

```auto

inputs = tokenizer(inputs, max_length=max_input_length, truncation=True, return_tensors="pt")
output = model.generate(**inputs, num_beams=8, do_sample=True, min_length=10, max_length=64)
decoded_output = tokenizer.batch_decode(output, skip_special_tokens=False)[0]
print(decoded_output)
predicted_Query = nltk.sent_tokenize(decoded_output.strip())[0]

print(predicted_Query)

```

the code above would output this:

```auto
<pad> db.movies.find(<unk>"title": "The Poor Little Rich Girl"<unk>, <unk>"writers": 1<unk>)</s>
<pad> db.movies.find(<unk>"title": "The Poor Little Rich Girl"<unk>, <unk>"writers": 1<unk>)</s>

```

The special token is supposed to be “{” or “}” . is there anyway to fix that ? also does that mean that during training it also outputted `<unk>` tokens?

Thanks for you time!
