Instructions to use guangyangnlp/Qwen3-4B-SFT-science-2e-5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use guangyangnlp/Qwen3-4B-SFT-science-2e-5 with Transformers:

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

pipe = pipeline("text-generation", model="guangyangnlp/Qwen3-4B-SFT-science-2e-5")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("guangyangnlp/Qwen3-4B-SFT-science-2e-5")
model = AutoModelForMultimodalLM.from_pretrained("guangyangnlp/Qwen3-4B-SFT-science-2e-5")
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 guangyangnlp/Qwen3-4B-SFT-science-2e-5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "guangyangnlp/Qwen3-4B-SFT-science-2e-5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "guangyangnlp/Qwen3-4B-SFT-science-2e-5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/guangyangnlp/Qwen3-4B-SFT-science-2e-5

SGLang

How to use guangyangnlp/Qwen3-4B-SFT-science-2e-5 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 "guangyangnlp/Qwen3-4B-SFT-science-2e-5" \
    --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": "guangyangnlp/Qwen3-4B-SFT-science-2e-5",
		"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 "guangyangnlp/Qwen3-4B-SFT-science-2e-5" \
        --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": "guangyangnlp/Qwen3-4B-SFT-science-2e-5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use guangyangnlp/Qwen3-4B-SFT-science-2e-5 with Docker Model Runner:
```
docker model run hf.co/guangyangnlp/Qwen3-4B-SFT-science-2e-5
```

Qwen3-4B-SFT-science-2e-5

This model is a fine-tuned version of Qwen/Qwen3-4B on the dolci_science_train dataset. It achieves the following results on the evaluation set:

Loss: 0.6771

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 16
total_train_batch_size: 128
total_eval_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.05
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.8054	0.2985	230	0.7241
0.6748	0.5969	460	0.7028
0.7003	0.8954	690	0.6880
0.5793	1.1933	920	0.6892
0.5997	1.4918	1150	0.6832
0.6137	1.7903	1380	0.6771
0.5071	2.0882	1610	0.6963
0.4680	2.3867	1840	0.6960
0.5263	2.6852	2070	0.6954
0.5248	2.9836	2300	0.6952

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 3

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for guangyangnlp/Qwen3-4B-SFT-science-2e-5

Base model

Qwen/Qwen3-4B-Base

Finetuned

Qwen/Qwen3-4B

Finetuned

(725)

this model