A lightweight retriever that sparsifies DeepSeek-V4 Compressed-Sparse-Attention (CSA) KV-cache.

Given the hidden state of a decode token, the retriever predicts which CSA KV-cache chunks the next ~64 tokens will attend to. Only the top-scoring chunks stay resident on the GPU; the rest can be offloaded to CPU/disk. In downstream evaluation it matches or beats the full-attention baseline while keeping ~10–15% of the KV cache on-device.

Model weights on Hugging Face

pip install torch safetensors

# Demo with mock inputs
python demo.py --ckpt weights/flashmemory_ds_v4.safetensors

# Toy sparse-decode loop
python toy_flashmemory_inference.py --ckpt weights/flashmemory_ds_v4.safetensors

from retriever import FlashMemoryRetriever

model = FlashMemoryRetriever.from_checkpoint(
    "weights/flashmemory_ds_v4.safetensors", device="cuda"
)

# hidden:     [B, 4096] decode-token hidden state
# comp_k:     [B, N, 132] uint8 compressed CSA keys
# positions:  [B] int64 token positions

# Per-layer sigmoid scores: {"l10": [B,N], "l12": [B,N], "l20": [B,N]}
per_layer = model(hidden, comp_k, positions)

# Cross-layer ensemble (mode="max" or "mean")
scores = model.ensemble(hidden, comp_k, positions, mode="max")       # [B, N]

# Boolean keep mask
keep = model.select_topk(hidden, comp_k, positions, top_k=512)       # top-K
keep = model.select_topk(hidden, comp_k, positions, threshold=0.5)   # threshold

Each chunk = HEAD_DIM + 4 = 132 uint8 bytes:

Dequant: fp8_values.view(float8_e4m3).float() * scale.
See make_mock_compressed_k() in demo.py.

Per CSA layer, scores are computed as:

hidden [B, 4096]
    → wq_a        (4096 → Q_LORA_RANK)
    → RMSNorm     (q_norm_weight, eps=1e-6)
    → wq_b        (Q_LORA_RANK → N_HEADS * HEAD_DIM)
    → reshape     [B, N_HEADS, HEAD_DIM]
    → RoPE        (YaRN, last ROPE_DIM=64 dims, base=160000)
    → Hadamard    (normalized Walsh-Hadamard)
    → q           [B, N_HEADS, HEAD_DIM]

hidden [B, 4096]
    → weights_proj (4096 → N_HEADS)
    → × weight_scale  (= HEAD_DIM^-0.5 * N_HEADS^-0.5)
    → fused_w     [B, N_HEADS]

compressed_k [B, N, HEAD_DIM + 4] (uint8)
    → bytes[:HEAD_DIM]  viewed as float8_e4m3 → dequant
    → × bytes[HEAD_DIM:]  viewed as float32   → k [B, N, HEAD_DIM]

score = sigmoid( sum_heads( relu(k @ q^T) * fused_w ) )   in [0, 1]

The checkpoint holds three independent CSA layers (l10, l12, l20), each with its own weights. At inference time per-layer sigmoid scores are ensembled per chunk — max (union, default) or mean — to produce a single keep/drop decision.

A self-contained illustration of how the retriever drives memory recall during decode — the actual control flow used inside DeepSeek-V4-FlashMemory.

 ┌──────────┐  compress & store    ┌────────────────────────────┐
 │ PREFILL  │  historical K/V      │  CSA KV-cache (the memory) │
 │ (dense   │ ───────────────────► │  N compressed chunks,      │
 │  attn)   │                      │  each = [132] uint8 fp8-K  │
 └────┬─────┘                      └──────────────┬─────────────┘
      │ last hidden state                         │ scored every 64 steps
      ▼                                           │
 ┌──────────────────────── DECODE LOOP ──────────┼──────────────────────────┐
 │ for each decode step t:                       │                          │
 │   hidden = toy_decoder.step(token, keep_mask)  │  (sparse memory attn)   │
 │                                               │                          │
 │   every RETRIEVAL_INTERVAL (= 64) steps:      ▼                          │
 │     scores[N]   = retriever.ensemble(hidden, compressed_k, pos)          │
 │     keep_mask[N] = top-K (or sigmoid>thresh) of scores                   │
 │     -> unselected chunks masked to -inf in next 64 steps                 │
 └──────────────────────────────────────────────────────────────────────────┘

1. Prefill (dense). Short prompt runs through dense memory attention. Its last hidden state seeds the first retrieval cycle.
2. Decode loop. Toy decoder produces a [B, 4096] hidden state each step.
3. Retrieval cycle (every 64 steps). The real FlashMemoryRetriever scores all N compressed-K chunks, ensembles per-layer scores, selects keep chunks.
4. Sparse attention. Unselected chunks' attention logits are set to -inf.

• This toy does NOT perform real CPU↔GPU KV-cache transfer. The swap engine is internal FlashMemory infrastructure and is not included.
• We simulate memory recall by masking attention logits to -inf. A masked chunk contributes nothing to attention — the same effect as not loading its KV.
• The purpose is to make the decode-time control flow concrete.

The production version depends on the internal sglang + DeepSeek-V4 CSA framework (native FP8 indexer, real compressed KV-cache, attention-sink, threshold fallback, per-request routing, actual KV swap) and cannot be released.

FlashMemory DS-V4 beats or ties the full-attention baseline on reasoning-heavy long-context tasks while keeping only ~10–15% of CSA KV cache on-device:

Precise needle-retrieval tasks (MRCR) require an additional threshold-fallback in the serving layer — this is not part of the standalone release.

MIT

If you use FlashMemory in your research, please cite:

@article{wang2026flashmemory,
  title   = {FlashMemory-DeepSeek-V4: Lightning Index Ultra-Long Context via Lookahead Sparse Attention},
  author  = {Yan Wang and Qifan Zhang and Jiachen Yu and Tian Liang and Dongyang Ma and
             Xiang Hu and Zibo Lin and Chunyang Li and Zhichao Wang and Jia Li and
             Yujiu Yang and Haitao Mi and Dong Yu},
  year    = {2026},
  journal = {arXiv preprint arXiv:2606.09079},
  url     = {https://huggingface.co/papers/2606.09079},
}