---
license: mit
tags:
- text-generation
- conversational
- falcon
- lora
- peft
- trl
- bitsandbytes
---

# Falcon-7b-chat-new-finetune

This model is a fine-tuned version of [ybelkada/falcon-7b-sharded-bf16](https://huggingface.co/ybelkada/falcon-7b-sharded-bf16) using the [timdettmers/openassistant-guanaco](https://huggingface.co/datasets/timdettmers/openassistant-guanaco) dataset. It has been fine-tuned using LoRA (Low-Rank Adaptation), PEFT (Parameter-Efficient Fine-Tuning), and TRL (Transformer Reinforcement Learning) techniques. It also leverages BitsAndBytes for 4-bit quantization.

## Model Description

This model is intended for conversational AI tasks, such as chatbots and dialogue systems. It has been trained on a large dataset of human conversations and is capable of generating human-like text.

## Intended Uses & Limitations

- **Intended uses:** This model can be used for text generation, dialogue systems, and other conversational AI applications.
- **Limitations:** The model may generate biased or offensive content. It is important to carefully review the model's outputs before using them in a production environment.

## Training and Fine-tuning

- **Base model:** [ybelkada/falcon-7b-sharded-bf16](https://huggingface.co/ybelkada/falcon-7b-sharded-bf16)
- **Dataset:** [timdettmers/openassistant-guanaco](https://huggingface.co/datasets/timdettmers/openassistant-guanaco)
- **Fine-tuning techniques:** LoRA, PEFT, TRL
- **Quantization:** BitsAndBytes (4-bit)

## How to Use
```
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

prompt = "Explain AI vs ML vs DL vs Generative AI."
model_name = "chaitanya42/Falcon-7b-chat-new-finetune"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

pipe = pipeline(task="text-generation", model=model, tokenizer=tokenizer, max_length=200)
result = pipe(f"{prompt} \n")
print(result[0]['generated_text'])
```

## Author

This model was fine-tuned by [chaitanya42](https://huggingface.co/chaitanya42).