Instructions to use hyper-accel/ci-random-qwen2-moe-a3b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hyper-accel/ci-random-qwen2-moe-a3b with Transformers:

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

pipe = pipeline("text-generation", model="hyper-accel/ci-random-qwen2-moe-a3b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("hyper-accel/ci-random-qwen2-moe-a3b")
model = AutoModelForCausalLM.from_pretrained("hyper-accel/ci-random-qwen2-moe-a3b")
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

vLLM

How to use hyper-accel/ci-random-qwen2-moe-a3b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "hyper-accel/ci-random-qwen2-moe-a3b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "hyper-accel/ci-random-qwen2-moe-a3b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/hyper-accel/ci-random-qwen2-moe-a3b

SGLang

How to use hyper-accel/ci-random-qwen2-moe-a3b 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 "hyper-accel/ci-random-qwen2-moe-a3b" \
    --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": "hyper-accel/ci-random-qwen2-moe-a3b",
		"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 "hyper-accel/ci-random-qwen2-moe-a3b" \
        --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": "hyper-accel/ci-random-qwen2-moe-a3b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use hyper-accel/ci-random-qwen2-moe-a3b with Docker Model Runner:
```
docker model run hf.co/hyper-accel/ci-random-qwen2-moe-a3b
```

ci-random-qwen2-moe-a3b / tokenizer.json

Commit History

Revert to e6fb385 — restore random init (distillation moved to hyper-accel/ci-random-qwen2-moe-a3b-distilled)

0f32a92
verified

ELutris commited on 7 days ago

KD-distilled 2-layer Qwen2-MoE student against Qwen1.5-MoE-A2.7B teacher (alpaca-cleaned, 1500 steps, T=2.0, KL loss)

ca82f0e
verified

ELutris commited on 7 days ago

Upload tokenizer from qwen2_moe

e6fb385
verified

junsoo999 commited on 15 days ago

Commit History

Revert to e6fb385 — restore random init (distillation moved to hyper-accel/ci-random-qwen2-moe-a3b-distilled) 0f32a92 verified

KD-distilled 2-layer Qwen2-MoE student against Qwen1.5-MoE-A2.7B teacher (alpaca-cleaned, 1500 steps, T=2.0, KL loss) ca82f0e verified

Upload tokenizer from qwen2_moe e6fb385 verified

Revert to e6fb385 — restore random init (distillation moved to hyper-accel/ci-random-qwen2-moe-a3b-distilled)

0f32a92
verified

KD-distilled 2-layer Qwen2-MoE student against Qwen1.5-MoE-A2.7B teacher (alpaca-cleaned, 1500 steps, T=2.0, KL loss)

ca82f0e
verified

Upload tokenizer from qwen2_moe

e6fb385
verified