Instructions to use jasonacox/jojo-124M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jasonacox/jojo-124M with Transformers:

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

pipe = pipeline("text-generation", model="jasonacox/jojo-124M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jasonacox/jojo-124M")
model = AutoModelForCausalLM.from_pretrained("jasonacox/jojo-124M")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use jasonacox/jojo-124M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jasonacox/jojo-124M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jasonacox/jojo-124M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/jasonacox/jojo-124M

SGLang

How to use jasonacox/jojo-124M with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "jasonacox/jojo-124M" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jasonacox/jojo-124M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "jasonacox/jojo-124M" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jasonacox/jojo-124M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use jasonacox/jojo-124M with Docker Model Runner:
```
docker model run hf.co/jasonacox/jojo-124M
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

jasonacox/jojo-124M

Model Description

jasonacox/jojo-124M is a GPT-style language model trained using the Jojo LLM training framework. This model was fine-tuned on the TinyStoriesV2 dataset and is designed for text generation tasks.

Model Details

Model Type: GPT-style Transformer Language Model
Training Framework: Jojo LLM
Language: English
License: MIT

Architecture

Layers: 12
Hidden Size: 768
Attention Heads: 12
Context Length: 1024 tokens
Vocabulary Size: 50,304 tokens
Total Parameters: 219.6M

Training Details

Training Data

The model was trained on the TinyStoriesV2 dataset.

Training Procedure

Training Framework: Jojo LLM v2.1.0
PyTorch Version: 2.7.1+cu126
Training Device: cuda:1
Precision: bfloat16

Training Hyperparameters

Batch Size: 12
Gradient Accumulation Steps: 40
Learning Rate: 0.0006
Weight Decay: 0.1
Dropout: 0.2
Gradient Clipping: 1.0

Training Results

Usage

Using with Transformers

from transformers import GPT2LMHeadModel, GPT2Tokenizer

# Load model and tokenizer
model = GPT2LMHeadModel.from_pretrained("jasonacox/jojo-124M")
tokenizer = GPT2Tokenizer.from_pretrained("gpt2")

# Generate text
input_text = "Your prompt here"
inputs = tokenizer.encode(input_text, return_tensors="pt")
outputs = model.generate(inputs, max_length=100, num_return_sequences=1, temperature=0.7)
generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(generated_text)

Using with Jojo LLM

# Generate text using the original Jojo framework
python gen.py jojo-124M.pt --prompt "Your prompt here"

Technical Specifications

Model Format: PyTorch
Precision: bfloat16
Framework Compatibility:
- ✅ Hugging Face Transformers
- ✅ Jojo LLM
- ✅ PyTorch

Model Card Authors

This model card was automatically generated by the Jojo LLM Hugging Face upload script.

Citation

If you use this model, please cite:

@misc{jasonacox/jojo_124m,
  title={jasonacox/jojo-124M},
  author={Jason A. Cox},
  year={2025},
  howpublished={\url{https://github.com/jasonacox/jojo}},
  note={Trained using Jojo LLM framework}
}

Framework Information

Jojo LLM Version: 2.1.0
Generation Date: 2025-07-10T22:23:12.633549
Checkpoint: jojo-124M.pt

For more information about the Jojo LLM framework, visit: https://github.com/jasonacox/jojo

Downloads last month: 8

Safetensors

Model size

0.1B params

Tensor type

F32

Model tree for jasonacox/jojo-124M

Quantizations

1 model

Evaluation results

Perplexity on Tinystoriesv2
self-reported

N/A