metadata
language:
- en
tags:
- gpt2
- text-generation
- story-generation
- creative-writing
- TinyStories
- ai-story
license: mit
datasets:
- roneneldan/TinyStories
metrics:
- perplexity
model-index:
- name: tiny-stories-gpt2
results: []
🧚♀️ Tiny Stories GPT-2
Author: Fathi7ma
Model type: Fine-tuned GPT-2 for creative story generation
Dataset: roneneldan/TinyStories
🧠 Model Overview
This model is a fine-tuned version of GPT-2 (small) trained on the TinyStories dataset — a large collection of short, simple stories for children.
The goal of this project is to create a lightweight, fun, and creative story generator that can produce short, imaginative stories from user prompts.
✨ Intended Use
Use this model for:
- Generating short, creative stories
- Educational or entertainment purposes
- Quick writing inspiration or kids’ content
Example use:
from transformers import pipeline
generator = pipeline("text-generation", model="Fathi7ma/tiny-stories-gpt2")
prompt = "A tiny robot who wanted to fly"
print(generator(prompt, max_length=80, num_return_sequences=1)[0]["generated_text"])
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.9127 | 1.0 | 2375 | 1.7789 |
| 1.7876 | 2.0 | 4750 | 1.7186 |
| 1.6976 | 3.0 | 7125 | 1.6995 |
### Framework versions
- Transformers 4.57.1
- Pytorch 2.8.0+cu126
- Datasets 4.0.0
- Tokenizers 0.22.1