Instructions to use lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound with Transformers:

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

pipe = pipeline("text-generation", model="lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound

SGLang

How to use lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound 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 "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound with Docker Model Runner:
```
docker model run hf.co/lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound
```

DeepSeek-V4.1-Flash — W4A16 (experts INT4 GPTQ + Engram INT4) — AutoRound

Post-training quantization of DeepSeek-V4.1-Flash produced with auto-round --model_free (RTN, iters=0):

Component	Official	This checkpoint
Routed experts (`ffn.experts.*`, incl. MTP)	MXFP4 (packed E2M1 + E8M0/32)	INT4 W4A16 g32 sym, auto_gptq packing (`qweight`/`qzeros`/`scales`)
Engram embedding tables (`layers.1/14.engram.embed`)	FP8 (E4M3 + E8M0/32)	INT4 g32 sym (packed uint8 `[R,128]` + fp16 scale `[R,8]`)
Attention / shared_experts / engram.wkv	FP8 / MXFP8	BF16 (dequantized)
Vision tower / aligner / embed / head / norms	BF16	BF16 (unchanged)

Checkpoint size: 451.7 GB (vs 510.3 GB official). gsm8k (raw 5-shot, thinking off, n=1319): 93.93 / 94.01 vs baseline 92.87 — lossless within noise.

Inference (vLLM)

Requires: vLLM main (≥ PR #56201) + the out-of-tree plugin dsv41-quant-plugin (registers the auto-round auto_gptq dispatch for deepseek_v41 and the INT4 engram lookup).

DSV41_ENGRAM_DTYPE=int4 NCCL_NVLS_ENABLE=0 \
vllm serve <this-model-dir> \
  --tensor-parallel-size 4 --max-model-len 8192 --language-model-only \
  --gpu-memory-utilization 0.90 --port 8100

Verified on 4×H200 (TP4). The INT4 expert GEMM runs through vLLM's built-in GPTQ-MoE path (MoeWNA16/Marlin) — no custom kernels; dense layers run as plain BF16.

Quantization recipe

# experts only -> auto_gptq INT4 g32 (everything else ignored / dequantized to bf16)
auto-round --model_name deepseek-ai/DeepSeek-V4.1-Flash --model_free \
  --scheme W4A16 --group_size 32 \
  --ignore_layers attn,shared_experts,engram,vision,aligner,main_proj,gate,norm,embed,head \
  --output_dir <A>

# engram tables -> int4 g32 sym + fp16 scale (custom RTN script), then splice into A
# (see producer notes: quant_engram_int4.py + build_b_view_experts.py)

Limitations

Serving today requires the dsv41-quant-plugin (not yet upstreamed to vLLM) and DSV41_ENGRAM_DTYPE=int4 for the engram lookup.
Not compatible with the official reference implementation's FP8 kernels (dense is BF16 here).

Provenance

auto-round 0.15.0 (editable copy with DeepSeek-V4.1 source-format patch), vLLM main nightly (0.1.1.dev39+g46d2b23ac), transformers-side loading not used (model_free).
Full production log & design docs: see producer's journal (aqa pipeline).

Downloads last month: 124

Safetensors

Model size

301B params

Tensor type

BF16

F32

I32

F16

Model tree for lvkaokao/DeepSeek-V4.1-Flash-W4A16-Engram-AutoRound

Base model

deepseek-ai/DeepSeek-V4.1-Flash

Quantized

(74)

this model