Instructions to use Ti-ger/Qwen1.5-4B-CQIA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Ti-ger/Qwen1.5-4B-CQIA with Transformers:

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

pipe = pipeline("text-generation", model="Ti-ger/Qwen1.5-4B-CQIA")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Ti-ger/Qwen1.5-4B-CQIA")
model = AutoModelForCausalLM.from_pretrained("Ti-ger/Qwen1.5-4B-CQIA")
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 Ti-ger/Qwen1.5-4B-CQIA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Ti-ger/Qwen1.5-4B-CQIA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Ti-ger/Qwen1.5-4B-CQIA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Ti-ger/Qwen1.5-4B-CQIA

SGLang

How to use Ti-ger/Qwen1.5-4B-CQIA 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 "Ti-ger/Qwen1.5-4B-CQIA" \
    --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": "Ti-ger/Qwen1.5-4B-CQIA",
		"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 "Ti-ger/Qwen1.5-4B-CQIA" \
        --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": "Ti-ger/Qwen1.5-4B-CQIA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Ti-ger/Qwen1.5-4B-CQIA with Docker Model Runner:
```
docker model run hf.co/Ti-ger/Qwen1.5-4B-CQIA
```

Model Card for Model ID

This model is fine-tuned on the Qwen1.5-4B-Chat model and COIG-CQIA/ruozhiba dataset by QLoRA.

Be noted that the model file contains the adapter LoRA weights only, so you are suggested to merge adapters with the base model for inference usage. Check the official reference here

This whole training process is running on Google Colab with free computing resources, detail can be accessed via link

This project is for the demonstration only of course DSAA5009, HKUST Guangzhou

Downloads last month: 6

Safetensors

Model size

4B params

Tensor type

F32

Ti-ger
/

Qwen1.5-4B-CQIA

Model Card for Model ID

Dataset used to train Ti-ger/Qwen1.5-4B-CQIA