README.md

---
license: apache-2.0
datasets:
- hakurei/open-instruct-v1
language:
- en
tags:
- code
- instruction-following
widget:
  - text: Tell me how to bake a cake
    example_title: Baking cakes
  - text: How can I print a fibonacci series upto N in C++
    example_title: Coding
---

# DialoGPT2 Instruction Following

This is the fine-tuned version of the [microsoft/dialogpt-small](https://huggingface.co/microsoft/DialoGPT-small) on the instruction following task. The dataset used was the [hakurei/open-instruct-v1](https://huggingface.co/datasets/hakurei/open-instruct-v1) dataset.

Find the training notebook here on [Kaggle](https://www.kaggle.com/code/smjishanulislam/basic-guide-on-instruction-following-dialogpt).


## Using the model


### Using `model.generate()`

To use the model, first call the checkpoints and initialize the model

```python
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("smji/dialogpt2-instruct-following")
model = AutoModelForCausalLM.from_pretrained("smji/dialogpt2-instruct-following")
```

And then move onto generating the text

```python
def generate_text(prompt):
    inputs = tokenizer.encode(prompt, return_tensors='pt').to(device)
    outputs = model.generate(inputs, max_length=512, pad_token_id=tokenizer.eos_token_id)
    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

    return generated_text[:generated_text.rfind('.')+1]

generate_text("How can I bake a cake?")
```

### Using the pipeline

Or, you can also use the pipeline

```python
# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="smji/dialogpt2-instruct-following")

pipe("How can I bake a cake?", max_length=512)
```

---

Done by [S M Jishanul Islam](https://github.com/S-M-J-I)