状态: 活跃研究
模型评估
准确率、鲁棒性、失效分析、领域迁移、泛化、可复现性,以及对局限的透明报告。
本方向的研究项目
- 状态: 实验性Apple Silicon 推理运行时
面向 Apple Silicon 及其他加速器上本地推理的内核设计、内存搬运与带宽上限研究——整合自早期 Octoryn Research 的工作。
- 状态: 实验性DAWM——深度感知世界模型
世界模型实验、阶段诊断与失效分析,把回归视为一连串被否证的假设,而非单一的训练失败。
开放问题
- RQ-001假设Apple GPU 能否支持持久化内核的全局同步?
- RQ-002证据收集中DAWM 第 7 阶段的哪些回归属于架构性问题,而非调参噪声?
发表
- Apple-Silicon-friendly LLM architecture: substrate laws reverse-engineered from a model bake-off
- Gemma4-26B-A4B on Apple Silicon: a drop-in isomorphism falsified (NO-GO), with a conditional new-port ceiling
- A falsify-first root cause for a concurrent 4-bit decode crash: batch-composition KV-pool wipe trips a re-seed precondition
- DiffusionGemma-26B compatibility assessment: NEEDS_ARCH_SUPPORT on Apple Silicon (pre-port, external llama.cpp baseline)
- Diffusion-LLM Decode Quality: An Honest Regression, Negative Evidence, and a Canvas-Geometry Reattribution
- Sovereign gpt-oss-20B (MXFP4) on Apple Silicon: a fine-grained-MoE small-matrix occupancy ceiling
- Per-stream KV-slot isolation for parallel low-bit decode on Apple Silicon: crash fix with byte-identical solo parity and no-regression
- Batched resident MoE experts on Apple Silicon: a flat parallel expert pass, ~1.6x decode, parity bit-identical (single-run)
- Batched-concurrent 4-bit serving on Apple Silicon: weight-stream amortization with an honest sublinear ceiling
- Amortizing the per-token host-sync setup tax in fused all-layer q4 decode on Apple M1 Ultra (1.57x), then occupancy-bound
- A sovereign ARM-NEON CPU inference backend (Raspberry Pi 4B): Q8 and multicore are complementary, not additive
- Independent GGUF Forward Pass and Bit-Identical Cross-Backend Consistency for a 26B Text-Diffusion Model (forward only, not full decode)
- Frozen-backbone slot binding exceeds feature clustering, with a characterized ceiling
- The Grounding Seam Failure: useful relations are not discovered from features
- Single-stream int4 decode of a 30B MoE on Apple Silicon: steady-state throughput with byte-identical, vendor-BLAS-free parity
- Resident packed-int4 decode: ~5x over host-streaming on a 24GB-class GPU
- FP8 grouped-expert prefill with on-chip int4-to-e4m3 dequant
- A DiT self-attention flash kernel on RDNA2: 9.4x op, 4.67x end-to-end
- DAWM: State over Token — a research charter with falsifiers
- An honest red result: EOS fixed, but forward degradation persisted
- ASR phase 1: a sovereign Whisper-tiny path landed
- DAWM State Contract: Observation to Event to Fluent
- Catalog of mechanism and boundary findings from a structured-state reasoning program
- A sovereign dense engine reproduces HF logits across 16 transformer families
- A deflated box-derived evaluation yardstick, and a (later-overturned) negative for frozen-backbone slot-binding
- Anti-collapse penalties transfer to real data, and a second collapse mode at the decoder in slot-based object discovery
- Porting Mamba (selective-scan SSM) into a sovereign sequence-mixing engine
- RWKV in a self-built engine: a second non-attention paradigm (WKV/RNN)
- Bloom in a sovereign engine: ALiBi as the third positional paradigm
- Falcon in a sovereign engine: the first parallel attention+MLP block
- Fully-sovereign text-to-video on Apple Silicon (Wan2.1), parity to diffusers
- OLMoE: the first sparse-MoE in a dense sovereign engine (and an eps trap)
- Granite: four scalar multipliers, and what a negative control cannot see
- OLMo-2: reordered (post) normalization and a loader-clobber root-cause
- StableLM-2: partial RoPE and a first LayerNorm-with-bias path in a dense ROCm engine
- Phi-3-mini: fused-tensor direct loading in a sovereign inference engine
- The direct HF-logits parity harness (and the compressed-tensors version trap)
- Continuous batching with paged KV: a dependency-free CUDA throughput path
- A multi-week debugging retrospective on an Apple Silicon 4-bit inference engine: command-buffer folding, an encoder leak, and cold-bank reservation
- True standalone text-to-video on gfx1030, sovereign (no vendor BLAS)
- True CUDA text-to-video with live UMT5 encode, dependency-minimal (no cuBLAS)
- 1024px image generation end-to-end on a consumer RDNA2 GPU, and a NULL-deref segfault root-caused
- Wan 3D-VAE decode parity on CPU (Swift): pinning the layout contract
- Loading image/video weights directly from HF safetensors (portability fix)
- Cross-Rendering Semantic Binding in an Engineered Small-Cell Schema (Mechanism Findings)
- DAWM Phase 3: selective persistence, contradiction handling, and learned abstain
- Soft cross-rendering convergence vs. graph-verified carrier-independence, and the value-erased-schema control boundary
- DAWM Phase 10: controlled small-cell integration coherence (preliminary)
- Controlled Symbolic Rendering-Invariance via Independent Learned Non-Text Parsers
- The slot-stationarity collapse boundary in learned variable-cardinality semantic schemas — a negative result
- Learned cursor routing under activation-graph cells: a negative-then-repair study of multi-candidate disambiguation, abstain stability, and deferred commitment
- DAWM Phase 9: cost-aware learned routing and budget-gated rollout
- DAWM Phase 7: multi-track consolidation
- Dual-channel reasoning state and a semantic-depth bridge for contradiction handling
- 论本地推理的实际边界:一则研究札记