Instructions to use LorMolf/SPSD-RL-Qwen3-4B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LorMolf/SPSD-RL-Qwen3-4B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="LorMolf/SPSD-RL-Qwen3-4B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("LorMolf/SPSD-RL-Qwen3-4B-Instruct")
model = AutoModelForMultimodalLM.from_pretrained("LorMolf/SPSD-RL-Qwen3-4B-Instruct")
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 LorMolf/SPSD-RL-Qwen3-4B-Instruct with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/LorMolf/SPSD-RL-Qwen3-4B-Instruct

SGLang

How to use LorMolf/SPSD-RL-Qwen3-4B-Instruct 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 "LorMolf/SPSD-RL-Qwen3-4B-Instruct" \
    --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": "LorMolf/SPSD-RL-Qwen3-4B-Instruct",
		"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 "LorMolf/SPSD-RL-Qwen3-4B-Instruct" \
        --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": "LorMolf/SPSD-RL-Qwen3-4B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use LorMolf/SPSD-RL-Qwen3-4B-Instruct with Docker Model Runner:
```
docker model run hf.co/LorMolf/SPSD-RL-Qwen3-4B-Instruct
```

SPSD-RL-Qwen3-4B-Instruct

This is the final loadable checkpoint from the Qwen3-4B-Instruct-2507 SDFT parity-400 run trained on SPSD-RL/MCTS-style supervision.

Local training artifact: outputs/qwen3-4b-instruct-2507-sdft-mctsstyle-parity400-lr1em6-cumem0-fixedfmt-20260611-194422

The trainer state in checkpoint-400 records global_step=400 and max_steps=400. Train-time evaluation was disabled after the step-100 TRL experimental SDFT eval-path crash; post-hoc reasoning and OpenReward benchmark evaluations are the source of truth for this artifact.

Uploaded Files

model.safetensors
config.json
generation_config.json
tokenizer.json
tokenizer_config.json
chat_template.jinja

Checkpoint directories, optimizer state, scheduler state, logs, local caches, and trainer process artifacts are intentionally excluded.

Evaluation Status

The fixed forced-boxed evaluation suite is launched separately from the local final model root using the repository sft_boxed_forced profile.

Downloads last month: 21

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for LorMolf/SPSD-RL-Qwen3-4B-Instruct

Base model

Qwen/Qwen3-4B-Instruct-2507

Finetuned

(1750)

this model