Instructions to use Josephgflowers/distillgpt2Cinder with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Josephgflowers/distillgpt2Cinder with Transformers:

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

pipe = pipeline("text-generation", model="Josephgflowers/distillgpt2Cinder")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Josephgflowers/distillgpt2Cinder")
model = AutoModelForCausalLM.from_pretrained("Josephgflowers/distillgpt2Cinder")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Josephgflowers/distillgpt2Cinder with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Josephgflowers/distillgpt2Cinder"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Josephgflowers/distillgpt2Cinder",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Josephgflowers/distillgpt2Cinder

SGLang

How to use Josephgflowers/distillgpt2Cinder 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 "Josephgflowers/distillgpt2Cinder" \
    --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": "Josephgflowers/distillgpt2Cinder",
		"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 "Josephgflowers/distillgpt2Cinder" \
        --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": "Josephgflowers/distillgpt2Cinder",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Josephgflowers/distillgpt2Cinder with Docker Model Runner:
```
docker model run hf.co/Josephgflowers/distillgpt2Cinder
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This is a distilgpt2 orginal run of Cinder with multi character chat. A little glitchy sometimes. The end sequence for each statement is <END>

If you would like multi chat with characters it is recommended to use <END>\nUSER: as the stop token.

USER: Hi, What is your name? Tell me about yourself.

CINDER: "I'm Cinder," a humanoid in my humanoid form. I was first introduced into the vastness of space during my lifetime.

Today, I'm a humanoid robot, and I'm designed for cosmic exploration, which is a subject of much speculation.

However, I'm designed for space exploration and scientific exploration.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	28.50
AI2 Reasoning Challenge (25-Shot)	24.49
HellaSwag (10-Shot)	27.24
MMLU (5-Shot)	24.97
TruthfulQA (0-shot)	43.96
Winogrande (5-shot)	50.12
GSM8k (5-shot)	0.23

Downloads last month: 68

Safetensors

Model size

81.9M params

Tensor type

F32

Model tree for Josephgflowers/distillgpt2Cinder

Quantizations

1 model

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

24.490
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

27.240
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

24.970
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

43.960
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

50.120
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

0.230