Instructions to use DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct")
model = AutoModelForCausalLM.from_pretrained("DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-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": "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct

SGLang

How to use DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-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 "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-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": "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-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 "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-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": "DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct with Docker Model Runner:
```
docker model run hf.co/DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct
```

DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B-Instruct on the sft_with_format dataset.

Model description

The model is the initial proposer of DERL, shared by all tasks. Earlier versions of the paper call it the meta-optimizer.

DERL is introduced in the paper Dual-Evolving Reinforcement Learning for Automated Reward Design.

Paper · Project page · Code

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 128
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 4
total_train_batch_size: 4096
total_eval_batch_size: 64
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 12.0

Training results

Framework versions

Transformers 4.52.4
Pytorch 2.6.0+cu124
Datasets 3.3.2
Tokenizers 0.21.1

Citation

@misc{cheng2026dualevolvingreinforcementlearningautomated,
      title={Dual-Evolving Reinforcement Learning for Automated Reward Design},
      author={Sitao Cheng and Tianle Li and Xuhan Huang and Xunjian Yin and Difan Zou},
      year={2026},
      eprint={2512.13399},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2512.13399},
}

Downloads last month: 504

Safetensors

Model size

0.5B params

Tensor type

BF16

Model tree for DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct

Base model

Qwen/Qwen2.5-0.5B

Finetuned

(758)

this model

Paper for DifferentiableEvolutionaryRL/DERL-Meta-Optimizer-Init-Qwen2.5-0.5B-Instruct

Differentiable Evolutionary Reinforcement Learning

Paper • 2512.13399 • Published Dec 15, 2025 • 22