Instructions to use SeongryongJung/Qwen3-4B-Physics-RLSD with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SeongryongJung/Qwen3-4B-Physics-RLSD with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD")
model = AutoModelForCausalLM.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD", 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 SeongryongJung/Qwen3-4B-Physics-RLSD with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/SeongryongJung/Qwen3-4B-Physics-RLSD

SGLang

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

Docker Model Runner
How to use SeongryongJung/Qwen3-4B-Physics-RLSD with Docker Model Runner:
```
docker model run hf.co/SeongryongJung/Qwen3-4B-Physics-RLSD
```

Qwen3-4B-Physics-RLSD / README.md

SeongryongJung

Upload Physics RLSD Qwen3-4B checkpoints

df44973 verified about 1 month ago

preview code

Raw

History Blame Contribute Delete

3.35 kB

metadata

license: apache-2.0
base_model: Qwen/Qwen3-4B
library_name: transformers
pipeline_tag: text-generation
tags:
  - qwen3
  - physics
  - reinforcement-learning
  - verl
  - sciknoweval
  - rlsd
  - self-distillation

Qwen3-4B Physics RLSD

This repository contains Physics fine-tuned Qwen3-4B checkpoints from the local SciKnowEval-style generalization setup.

Root checkpoint: final global_step_100 merged to Hugging Face safetensors.
best_avg16/: checkpoint with the highest validation avg@16 during training, merged to Hugging Face safetensors.

Checkpoints

Checkpoint	Source step	Validation avg@16	best@16 / pass@16	maj@16
Root final	100	0.700000	0.768713	0.718600
`best_avg16/`	60	0.731250	0.816338	0.740938

Training Run

qwen3gen-physics-RLSD-Qwen-Qwen3-4B-mbs8-decay0-ema0.05-train256-rollout8-lr1e-6-vllm0.8

Base Model

Base model: Qwen/Qwen3-4B
Fine-tuning type: full-parameter FSDP RL training
Dataset: datasets/sciknoweval/physics
Train split: 720 examples
Validation split: 80 examples

Method

Method: RLSD
Config: rlsd
Policy loss mode: rlsd
Reward: local SciKnowEval multiple-choice reward checker
Rollout correction: token-level importance sampling, threshold 2.0

Hyperparameters

Field	Value
Base model	`Qwen/Qwen3-4B`
Training steps	100
Train batch size	256
Rollouts per prompt	8
Generations per step	2048
PPO mini batch size	8
Learning rate	`1e-6`
LR warmup steps	10
Weight decay	0.01
Grad clip	1.0
Max prompt length	2048
Max response length	8192
Max model length	10240
Train temperature	1.0
Train top_p	1.0
Validation generations	16
Validation temperature	0.6
Validation top_p	0.95
vLLM GPU memory utilization	0.8
GPUs	8 x NVIDIA H200
Save frequency	every 10 steps
Validation frequency	every 10 steps
Token reweight lambda	0.5
Token reweight eps_w	0.2
Token reweight decay steps	0
Teacher update rate	0.05
Max reprompt length	10240

Metrics

Metric	Value
Final training step	100
Final `critic/score/mean`	0.890137
Final `critic/rewards/mean`	0.890137
Final validation `avg@16`	0.700000
Peak validation `avg@16`	0.731250
Peak validation step	60

Loading

Root final checkpoint:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD")
tokenizer = AutoTokenizer.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD")

Best avg@16 checkpoint:

model = AutoModelForCausalLM.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD", subfolder="best_avg16")
tokenizer = AutoTokenizer.from_pretrained("SeongryongJung/Qwen3-4B-Physics-RLSD", subfolder="best_avg16")

Intended Use

This model is intended for research on RL fine-tuning and self-distillation behavior on science/generalization tasks. It has not been broadly safety evaluated for production use.

Limitations

The reported scores are training-time and validation-time metrics from the local experimental setup. They should not be interpreted as broad benchmark results without independent evaluation.