TinyStories: How Small Can Language Models Be and Still Speak Coherent English?
Paper • 2305.07759 • Published • 50
How to use evanhugging/tiny-stories-50m-GGUF with llama.cpp:
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf evanhugging/tiny-stories-50m-GGUF:F16 # Run inference directly in the terminal: llama cli -hf evanhugging/tiny-stories-50m-GGUF:F16
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf evanhugging/tiny-stories-50m-GGUF:F16 # Run inference directly in the terminal: llama cli -hf evanhugging/tiny-stories-50m-GGUF:F16
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf evanhugging/tiny-stories-50m-GGUF:F16 # Run inference directly in the terminal: ./build/bin/llama cli -hf evanhugging/tiny-stories-50m-GGUF:F16
docker model run hf.co/evanhugging/tiny-stories-50m-GGUF:F16
How to use evanhugging/tiny-stories-50m-GGUF with Ollama:
ollama run hf.co/evanhugging/tiny-stories-50m-GGUF:F16
How to use evanhugging/tiny-stories-50m-GGUF with Docker Model Runner:
docker model run hf.co/evanhugging/tiny-stories-50m-GGUF:F16
How to use evanhugging/tiny-stories-50m-GGUF with Lemonade:
# Download Lemonade from https://lemonade-server.ai/ lemonade pull evanhugging/tiny-stories-50m-GGUF:F16
lemonade run user.tiny-stories-50m-GGUF-F16
lemonade list
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
I trained a 51M parameter GPT-style model on an RTX 3080 10GB using TinyStories. It can generate coherent children’s-story style text. Includes HF Transformers weights and GGUF/Ollama export. Followed https://arxiv.org/abs/2305.07759
16-bit