Qwen3 4B DFlash b16

uv pip install "git+https://github.com/sgl-project/sglang.git@refs/pull/20547/head#subdirectory=python"

# Optional: enable schedule overlapping (experimental, may not be stable)
# export SGLANG_ENABLE_SPEC_V2=1
# export SGLANG_ENABLE_DFLASH_SPEC_V2=1
# export SGLANG_ENABLE_OVERLAP_PLAN_STREAM=1

python -m sglang.launch_server \
    --model-path Qwen/Qwen3-4B \
    --speculative-algorithm DFLASH \
    --speculative-draft-model-path z-lab/Qwen3-4B-DFlash-b16 \
    --tp-size 1 \
    --dtype bfloat16 \
    --attention-backend fa3 \
    --mem-fraction-static 0.75 \
    --trust-remote-code

from openai import OpenAI

client = OpenAI(base_url="http://localhost:30000/v1", api_key="EMPTY")

response = client.chat.completions.create(
    model="Qwen/Qwen3-4B",
    messages=[{"role": "user", "content": "Write a quicksort in Python."}],
    max_tokens=2048,
    temperature=0.0,
    extra_body={
        "chat_template_kwargs": {"enable_thinking": False},
    },
)
print(response.choices[0].message.content)

uv pip install vllm
uv pip install -U vllm --torch-backend=auto --extra-index-url https://wheels.vllm.ai/nightly

vllm serve Qwen/Qwen3-4B \
  --speculative-config '{"method": "dflash", "model": "z-lab/Qwen3-4B-DFlash-b16", "num_speculative_tokens": 15}' \
  --attention-backend flash_attn \
  --max-num-batched-tokens 32768

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")

response = client.chat.completions.create(
    model="Qwen/Qwen3-4B",
    messages=[{"role": "user", "content": "Write a quicksort in Python."}],
    max_tokens=2048,
    temperature=0.0,
    chat_template_kwargs: {"enable_thinking": False},
)
print(response.choices[0].message.content)

pip install transformers==4.57.3 torch==2.9.1 accelerate

from transformers import AutoModel, AutoModelForCausalLM, AutoTokenizer

# 1. Load the DFlash Draft Model
# Note: trust_remote_code=True is required for DFlash. We recommend run on one GPU currently.
model = AutoModel.from_pretrained(
    "z-lab/Qwen3-4B-DFlash-b16", 
    trust_remote_code=True, 
    dtype="auto", 
    device_map="cuda:0"
).eval()

# 2. Load the Target Model
target = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen3-4B", 
    dtype="auto", 
    device_map="cuda:0"
).eval()

# 3. Load Tokenizer and Prepare Input
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3-4B")
prompt = "How many positive whole-number divisors does 196 have?"
messages = [
    {"role": "user", "content": prompt}
]
# Note: this draft model is used for thinking mode disabled
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=False
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# 4. Run Speculative Decoding
# The 'spec_generate' function is a custom method provided by the DFlash model
generate_ids = model.spec_generate(
    input_ids=model_inputs["input_ids"], 
    max_new_tokens=2048, 
    temperature=0.0, 
    target=target, 
    stop_token_ids=[tokenizer.eos_token_id]
)

print(tokenizer.decode(generate_ids[0], skip_special_tokens=True))

@misc{chen2026dflash,
  title         = {DFlash: Block Diffusion for Flash Speculative Decoding},
  author        = {Chen, Jian and Liang, Yesheng and Liu, Zhijian},
  year          = {2026},
  eprint        = {2602.06036},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2602.06036}
}