Instructions to use WyvernAI/gpt2-scratch-small-ftn-baseline with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use WyvernAI/gpt2-scratch-small-ftn-baseline with Transformers:

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

pipe = pipeline("text-generation", model="WyvernAI/gpt2-scratch-small-ftn-baseline")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("WyvernAI/gpt2-scratch-small-ftn-baseline")
model = AutoModelForCausalLM.from_pretrained("WyvernAI/gpt2-scratch-small-ftn-baseline")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use WyvernAI/gpt2-scratch-small-ftn-baseline with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "WyvernAI/gpt2-scratch-small-ftn-baseline"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "WyvernAI/gpt2-scratch-small-ftn-baseline",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/WyvernAI/gpt2-scratch-small-ftn-baseline

SGLang

How to use WyvernAI/gpt2-scratch-small-ftn-baseline 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 "WyvernAI/gpt2-scratch-small-ftn-baseline" \
    --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": "WyvernAI/gpt2-scratch-small-ftn-baseline",
		"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 "WyvernAI/gpt2-scratch-small-ftn-baseline" \
        --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": "WyvernAI/gpt2-scratch-small-ftn-baseline",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use WyvernAI/gpt2-scratch-small-ftn-baseline with Docker Model Runner:
```
docker model run hf.co/WyvernAI/gpt2-scratch-small-ftn-baseline
```

GPT-2 Scratch Small FTN Baseline

Scratch-trained small GPT-2 baseline used for FTN comparisons on TinyStories.

This checkpoint is the small scratch GPT-2 baseline used for FTN comparison experiments.

Training summary

Layers: 4
Hidden size: 256
Heads: 4
Context length: 256
Best validation loss: 3.590700
Best epoch: 10

Files

standard Hugging Face GPT-2 weights and config
metrics_summary.json
metrics_history.csv
samples.json / samples.txt

Load the model

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained('.')
tokenizer = AutoTokenizer.from_pretrained('.')

Downloads last month: 4

Safetensors

Model size

16.1M params

Tensor type

F32