Instructions to use junnyu/DeepScaleR-1.5B-Preview-Reproduce with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use junnyu/DeepScaleR-1.5B-Preview-Reproduce with Transformers:

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

pipe = pipeline("text-generation", model="junnyu/DeepScaleR-1.5B-Preview-Reproduce")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("junnyu/DeepScaleR-1.5B-Preview-Reproduce")
model = AutoModelForCausalLM.from_pretrained("junnyu/DeepScaleR-1.5B-Preview-Reproduce", 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 junnyu/DeepScaleR-1.5B-Preview-Reproduce with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "junnyu/DeepScaleR-1.5B-Preview-Reproduce"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "junnyu/DeepScaleR-1.5B-Preview-Reproduce",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/junnyu/DeepScaleR-1.5B-Preview-Reproduce

SGLang

How to use junnyu/DeepScaleR-1.5B-Preview-Reproduce 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 "junnyu/DeepScaleR-1.5B-Preview-Reproduce" \
    --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": "junnyu/DeepScaleR-1.5B-Preview-Reproduce",
		"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 "junnyu/DeepScaleR-1.5B-Preview-Reproduce" \
        --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": "junnyu/DeepScaleR-1.5B-Preview-Reproduce",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use junnyu/DeepScaleR-1.5B-Preview-Reproduce with Docker Model Runner:
```
docker model run hf.co/junnyu/DeepScaleR-1.5B-Preview-Reproduce
```

DeepScaleR-1.5B-Preview-Reproduce

File size: 2,533 Bytes

933246b

#!/bin/bash
set -x

# Warning: Export VLLM_ATTENTION_BACKEND on every machine before starting Ray cluster.
# vLLM without XFORMERS will results in CUDA errors.
export VLLM_ATTENTION_BACKEND=XFORMERS

# Parse command line arguments
while [[ $# -gt 0 ]]; do
    case $1 in
        --model)
            MODEL_PATH="$2"
            shift 2
            ;;
        *)
            break
            ;;
    esac
done

# Check if model path is provided
if [ -z "$MODEL_PATH" ]; then
    MODEL_PATH="deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B"
fi

# Train over 4 nodes, 8 A100-80GB GPUs per node.
python3 -m verl.trainer.main_ppo \
    algorithm.adv_estimator=grpo \
    data.train_files=$HOME/deepscaler/data/train.parquet \
    data.val_files=$HOME/deepscaler/data/aime.parquet \
    data.train_batch_size=64 \
    data.val_batch_size=128 \
    data.max_prompt_length=1024 \
    data.max_response_length=24576 \
    actor_rollout_ref.model.path=$MODEL_PATH \
    actor_rollout_ref.actor.optim.lr=1e-6 \
    actor_rollout_ref.model.use_remove_padding=True \
    actor_rollout_ref.actor.ppo_mini_batch_size=64 \
    actor_rollout_ref.actor.ppo_epochs=1 \
    actor_rollout_ref.actor.use_dynamic_bsz=True \
    actor_rollout_ref.actor.ppo_max_token_len_per_gpu=32768 \
    actor_rollout_ref.actor.use_kl_loss=True \
    actor_rollout_ref.actor.kl_loss_coef=0.001 \
    actor_rollout_ref.actor.kl_loss_type=low_var_kl \
    actor_rollout_ref.actor.ulysses_sequence_parallel_size=1 \
    actor_rollout_ref.model.enable_gradient_checkpointing=True \
    actor_rollout_ref.actor.fsdp_config.param_offload=False \
    actor_rollout_ref.actor.fsdp_config.grad_offload=False \
    actor_rollout_ref.actor.fsdp_config.optimizer_offload=False \
    actor_rollout_ref.rollout.tensor_model_parallel_size=1 \
    actor_rollout_ref.rollout.name=vllm \
    actor_rollout_ref.rollout.temperature=0.6 \
    actor_rollout_ref.rollout.val_temperature=0.6 \
    actor_rollout_ref.rollout.gpu_memory_utilization=0.7 \
    actor_rollout_ref.rollout.n=16 \
    actor_rollout_ref.rollout.n_val=16 \
    actor_rollout_ref.ref.fsdp_config.param_offload=True \
    algorithm.kl_ctrl.kl_coef=0.001 \
    trainer.critic_warmup=0 \
    trainer.logger=['wandb'] \
    trainer.project_name='deepscaler' \
    trainer.experiment_name='deepscaler-1.5b-24k' \
    +trainer.val_before_train=False \
    trainer.n_gpus_per_node=8 \
    trainer.nnodes=1 \
    trainer.save_freq=10 \
    trainer.test_freq=10 \
    trainer.default_hdfs_dir=null \
    trainer.total_epochs=30 "${@:1}"