Instructions to use lolxdasfwa/Quartz-S1-Base-33.89M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lolxdasfwa/Quartz-S1-Base-33.89M with Transformers:

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

pipe = pipeline("text-generation", model="lolxdasfwa/Quartz-S1-Base-33.89M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("lolxdasfwa/Quartz-S1-Base-33.89M")
model = AutoModelForCausalLM.from_pretrained("lolxdasfwa/Quartz-S1-Base-33.89M", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use lolxdasfwa/Quartz-S1-Base-33.89M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lolxdasfwa/Quartz-S1-Base-33.89M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lolxdasfwa/Quartz-S1-Base-33.89M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/lolxdasfwa/Quartz-S1-Base-33.89M

SGLang

How to use lolxdasfwa/Quartz-S1-Base-33.89M 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 "lolxdasfwa/Quartz-S1-Base-33.89M" \
    --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": "lolxdasfwa/Quartz-S1-Base-33.89M",
		"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 "lolxdasfwa/Quartz-S1-Base-33.89M" \
        --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": "lolxdasfwa/Quartz-S1-Base-33.89M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use lolxdasfwa/Quartz-S1-Base-33.89M with Docker Model Runner:
```
docker model run hf.co/lolxdasfwa/Quartz-S1-Base-33.89M
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

About

Quartz-S1-Base-33.89M is a pre-trained AI model ready to be fine tuned (or use it as it is lol)

It uses GPT-2 architecture

Trained on: HuggingFaceFW/fineweb-edu

Training

Trained for 27,000 steps

Final Loss: 5.313 Final validation loss: 5.321

List of Features

Text Generation (e.g., trying to complecte text)
Fine Tuning Capabilities (e.g., customizing the model)
Small Language Model Capabilities (e.g., while it is very dumb once you fine tune it it can answer basic questions)

Summary

It is a very small language model (SLM) and has very high chances to produce incorrect, repetitive, or nonsensical text. It should not be expected to have the capabilities of much larger language models.

Downloads last month: 660

Safetensors

Model size

33.9M params

Tensor type

F32

Model tree for lolxdasfwa/Quartz-S1-Base-33.89M

Unable to build the model tree, the base model loops to the model itself. Learn more.

lolxdasfwa
/

Quartz-S1-Base-33.89M

About

Training

List of Features

Summary

Model tree for lolxdasfwa/Quartz-S1-Base-33.89M

Dataset used to train lolxdasfwa/Quartz-S1-Base-33.89M