Recipe
deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2
deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2Historically accepted DeepSeek-V4-Flash-DSpark NVFP4 across two DGX Spark nodes with TP2, DSpark speculative decoding, NVFP4 DS-MLA KV, and CUDA graphs; awaiting fresh two-node acceptance
Record
- Status
- candidate
- Source
- 0xsero
- Engine
- vllm
- Engine version
- 0.25.2+anemll-dspark-gx10-0.1.1
- Graph
- full-and-piecewise
- Accelerators
- 2
- Tensor parallel
- 2
- Max concurrency
- 6
- chat
- yes
- reasoning
- yes
- tools
- yes
- vision
- no
Hugging Face model card
Identity
https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-DSpark- Repository
- deepseek-ai/DeepSeek-V4-Flash-DSpark
- Status
- known
- Link type
- Exact Hub repository
Public Hugging Face repository confirmed by the Hub API.
Candidate: useful compatibility or speed evidence. The registry does not offer Run until promotion requirements are met.
Compose configuration
vllm
Container · candidate · docker compose
Candidate evidence — not a Run contract
- Compose file
compose.yml- Port
- 8888
Environment
| Variable | Value |
|---|---|
CUTE_DSL_ARCH | sm_121a |
FLASHINFER_CUDA_ARCH_LIST | 12.1a |
GPU_MEMORY_UTILIZATION | 0.80 |
KV_CACHE_DTYPE | nvfp4_ds_mla |
MAX_MODEL_LEN | 1048576 |
MAX_NUM_BATCHED_TOKENS | 8192 |
MAX_NUM_SEQS | 6 |
MTP_NUM_TOKENS | 5 |
NCCL_IB_DISABLE | 0 |
NCCL_NET | IB |
TORCH_CUDA_ARCH_LIST | 12.1a |
VLLM_USE_B12X_MOE | 1 |
VLLM_USE_FLASHINFER_SAMPLER | 1 |
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status | Sweep |
|---|---|---|---|---|---|---|
| 1 | — | — | 66.6 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
| 2 | — | — | 47.2 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
| 3 | — | — | 31.9 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
| 4 | — | — | 32.8 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
| 5 | — | — | 25.7 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
| 6 | — | — | 26.8 | — | historical | deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- hardware count
- 2
- hardware id
- dgx-spark-gb10-128gb
- id
- deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2
- model instance id
- deepseek-ai-deepseek-v4-flash-dspark--nvfp4
- recipe source
- 0xsero
- schema version
- local-ai-registry/v1
- speed sweep ids
- deepseek-v4-flash-dspark-nvfp4-dgxspark-vllm-tp2-sweep
- status
- candidate
capabilities
- chat
- Yes
- reasoning
- Yes
- tools
- Yes
- vision
- No
engine
- graph mode
- full-and-piecewise
- name
- vllm
- version
- 0.25.2+anemll-dspark-gx10-0.1.1
serving
- max concurrency
- 6
- tensor parallel
- 2
Provenance & metadata (3)
facts
- serving.kv cache tokens · provenance · captured at
- 2026-08-27T06:04:13.773Z
serving.kv cache tokens · reason kv-cache-capacity-not-published
serving.kv cache tokens · state unknown
- serving.max context tokens · provenance · captured at
- 2026-08-27T06:04:13.773Z
serving.max context tokens · reason not-observed
serving.max context tokens · state unknown
metadata
provenance
- captured at
- 2026-08-27T06:04:13.773Z
sources
| captured at | kind | url |
|---|---|---|
| 2026-08-27T06:04:13.773Z | normalized-recipe | github.com/0xSero/local-ai-registry ↗ |