Instructions to use raincandy-u/Quark-464M-v0.1.alpha with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use raincandy-u/Quark-464M-v0.1.alpha with Transformers:

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

pipe = pipeline("text-generation", model="raincandy-u/Quark-464M-v0.1.alpha")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("raincandy-u/Quark-464M-v0.1.alpha")
model = AutoModelForCausalLM.from_pretrained("raincandy-u/Quark-464M-v0.1.alpha", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use raincandy-u/Quark-464M-v0.1.alpha with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "raincandy-u/Quark-464M-v0.1.alpha"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "raincandy-u/Quark-464M-v0.1.alpha",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/raincandy-u/Quark-464M-v0.1.alpha

SGLang

How to use raincandy-u/Quark-464M-v0.1.alpha 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 "raincandy-u/Quark-464M-v0.1.alpha" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "raincandy-u/Quark-464M-v0.1.alpha",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "raincandy-u/Quark-464M-v0.1.alpha" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "raincandy-u/Quark-464M-v0.1.alpha",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use raincandy-u/Quark-464M-v0.1.alpha with Docker Model Runner:
```
docker model run hf.co/raincandy-u/Quark-464M-v0.1.alpha
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🌟⚛ Introducing Quark Series: Empowering Edge Devices with Swift Bilingual Conversational AI

Presenting Quark-620M-v0.1.alpha, the first model in our Quark series.

Quark models focus on delivering exceptional English and Chinese conversational performance on edge devices with rapid inference speed.

🗨 Example

🧑‍🏫 Benchmark

⏳Wait for uploading

🔍 Disclaimer

As an alpha preview release without RLHF fine-tuning, we do not take responsibility for potentially harmful responses and are committed to continuous improvement based on user feedback and research.

👥 Join Our Discord community!

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	35.68
AI2 Reasoning Challenge (25-Shot)	31.40
HellaSwag (10-Shot)	47.31
MMLU (5-Shot)	34.55
TruthfulQA (0-shot)	41.84
Winogrande (5-shot)	55.17
GSM8k (5-shot)	3.79

Downloads last month: 24

Safetensors

Model size

0.5B params

Tensor type

F32

Model tree for raincandy-u/Quark-464M-v0.1.alpha

Quantizations

2 models

Dataset used to train raincandy-u/Quark-464M-v0.1.alpha

Evaluation results

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

31.400
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

47.310
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

34.550
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

41.840
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

55.170
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

3.790