hf2q

Pure-Rust CLI for converting HuggingFace models to hardware-optimized formats — and serving them through an OpenAI-compatible HTTP API on Apple Silicon. No C++ at build, test, or runtime (ADR-008 sovereignty rule); the inference path runs entirely on mlx-native Metal kernels we own end-to-end.

Performance — on M5 Max at HEAD (2026-05-17 re-bench, 3-run median, default config including the HF2Q_NO_FA hybrid-attn fix from commit 03328ee5):

• Gemma-4 26B-A4B Q6_K decode — tg200 105.2 t/s vs llama.cpp -fa 1 104.32 t/s (1.01× peer-FA AHEAD); tg2000 93.5 t/s vs 96.69 t/s (0.97× peer-FA).
• Qwen 3.6 35B-A3B APEX-Q5_K_M decode (TQ-V default-on) — tg200 130.6 t/s vs llama.cpp -fa 1 100.97 t/s (1.29× peer-FA AHEAD); tg1500 129.1 t/s vs 89.25 t/s (1.45× peer-FA AHEAD — TQ-V's bandwidth advantage widens with depth). Byte-identical to llama.cpp for the first 242 bytes of greedy output (sourdough_qwen35.sh gate).
• Gemma-4 prefill — pp1800 2734 t/s vs llama.cpp 2837 t/s (0.96× peer-FA); pp3700 2703 t/s vs 2181 t/s (1.24× peer-FA AHEAD) — hf2q's prefill rate drops only ~1% from pp1800→pp3700 while llama's drops ~23%, so the cross-over is in the lower part of this range.
• TurboQuant 8-bit KV cache — Qwen 3.6 35B-A3B at 32K context: 340 MiB vs 1.34 GiB F32 baseline = 3.94× memory savings (ADR-027 iter-34, regression-pinned by tests/qh35_no_f32_kv_alloc_with_tq_kv.rs). Default-on for Qwen 3.5/3.6 as of 2026-05-17 — opt out with HF2Q_TQ_KV=0.
• DeepSeek-V4-Flash-0731 Q2_K_S decode — 45.1 t/s median on the official 89.65 GiB hf2q-converted artifact vs 41.58 t/s for the pinned llama.cpp reference. Required-tool grammar decode reaches 25.4 t/s and returns structured OpenAI tool calls.

Methodology references in docs/peer-parity-baselines-2026-04-26.md; the historical 1.05× decode + 1.07-1.09× prefill claims (ADR-029 iter-175) were measured at a pre-HF2Q_NO_FA HEAD and do not hold at current main per the re-bench above.

# Convert a HuggingFace model to a Q4_K_M GGUF (auto-downloads via --repo)
hf2q convert \
  --repo google/gemma-4-26b-it \
  --quant q4_k_m \
  --output models/gemma-4-26b-it-q4_k_m/out.gguf

# Serve it over an OpenAI-compatible HTTP API
hf2q serve --model models/gemma-4-26b-it-q4_k_m/out.gguf --port 8080

What it does

hf2q is two tools fused into one binary:

1. A conversion pipeline. Read HuggingFace config.json + *.safetensors, normalize tensor names per architecture, run quantization (legacy block Q4_0 / Q8_0, K-quants Q{2..6}_K_{S,M,L}, imatrix-weighted K-quants including imatrix-adaptive, or mixed-bit dynamic-quant-*) and emit GGUF or mlx-lm safetensors. No llama.cpp or candle is involved at build, test or runtime (ADR-008 — "candle divorce"; sovereignty rule in docs/arch-onboarding.md).

2. An inference + serving engine. Load a GGUF, run prefill + speculative-or-vanilla decode on the GPU via mlx-native, expose it through an OpenAI-style /v1/chat/completions, /v1/embeddings and /v1/models HTTP API. Supports tools / function-calling, streaming SSE, vision (qwen3vl), grammar-constrained sampling, and a persistent block-prefix KV cache.

Supported architectures today: Gemma 4 (dense + MoE), Qwen 3.5 / 3.6 (dense + MoE + multi-token-prediction), DeepSeek-V4-Flash-0731 (compressed-attention MoE), Qwen 3-VL (vision + text), and BERT / Nomic-BERT (embedding-only). Each lives under a single src/inference/models/<arch>/ module — the arch-registry (src/arch/) is the single source of truth for tensor catalogs, quality thresholds, smoke prompts and MTP/vision flags.

Install

hf2q is a Cargo crate. Apple Silicon is currently the only supported target — the inference path is Metal-only.

git clone git@github.com:robertelee78/hf2q.git
cd hf2q
cargo build --release
./target/release/hf2q --help

The default mlx-native = "0.9" declaration at Cargo.toml:105 resolves from crates.io. For local mlx-native development place a path override in a gitignored .cargo/config.toml (template at Cargo.toml:217+) — out-of-the-box cargo build does NOT path-pin to a sibling checkout.

cargo build requires:

• macOS with Metal Performance Shaders (M1 or newer).
• A working Rust toolchain at the version pinned in Cargo.toml (rust-version = "1.81.0").
• Per-arch disk floor for convert (src/arch/entries/): 100 GB for Qwen 3.5 dense, 150 GB for Qwen 3.5 MoE. Smoke preflight refuses to start below disk_floor_gb + 10.

hf2q doctor enumerates the runtime checks (hardware detection, disk space, optional RuVector backend); run it after cargo install if anything misbehaves.

CLI subcommands

Run hf2q <command> --help for the full flag surface.

Quantization variants

The hf2q convert pipeline accepts two families of --quant <name> values, parsed via QuantSelector::from_name:

Reserved names surface as typed errors with actionable hints: --quant dwq → "reserved for the future DWQ-train pipeline"; --quant apex (unqualified) → suggests apex-balanced etc.; --quant tq1_0/tq2_0 → "recognized ftype but out of v1 scope".

Quick start: convert + serve a model

The hf2q convert pipeline reads a HuggingFace model directory (config.json + safetensors + tokenizer.json) and emits a single GGUF that loads in stock llama.cpp and in hf2q serve. The source can be a path that already exists on disk OR a --repo <hf_repo> that the driver auto-downloads via huggingface-cli.

# 1. Pre-download the HF source explicitly:
huggingface-cli download google/gemma-4-26b-a4b-it \
  --local-dir ./models/google-gemma-4-26b-a4b-it

# 2. Convert to Q5_K_M. Streaming convert keeps peak memory ~5 GB
#    even on a 48 GB-source 26 B-param model. ~8-15 min on M-series.
hf2q convert ./models/google-gemma-4-26b-a4b-it \
  --quant q5_k_m \
  -o ./out/gemma4-26b-q5_k_m.gguf

# Alternative: --repo auto-downloads via huggingface-cli into
# ~/.cache/hf2q/repos/google__gemma-4-26b-a4b-it/ and then converts.
# Mutually exclusive with the positional path form above.
hf2q convert --repo google/gemma-4-26b-a4b-it \
  --quant q5_k_m \
  -o ./out/gemma4-26b-q5_k_m.gguf

# 3a. Test load with stock llama.cpp (single-shot generation):
llama-cli -m ./out/gemma4-26b-q5_k_m.gguf \
  -p "What is the capital of France?" -n 64 --temp 0 --seed 42

# 3b. Serve with hf2q's OpenAI-compatible HTTP API:
hf2q serve --model ./out/gemma4-26b-q5_k_m.gguf --port 8080

# 4. Use it (OpenAI SDK works out of the box)
curl -X POST http://localhost:8080/v1/chat/completions \
  -H 'Content-Type: application/json' \
  -d '{"model":"gemma4","messages":[{"role":"user","content":"hello"}]}'

Full-context agentic serving

The native Gemma 4, Qwen 3.6, and DeepSeek-V4 workers are intended for OpenAI-compatible coding clients such as OpenCode. Their canonical launchers default to four independent agent slots. Every slot receives the complete configured logical context; model weights are shared, while KV, recurrent state, token ledgers, template state, and tool-call state remain isolated per conversation. One shared physical KV budget governs demand-grown residency—it never divides the advertised context by the slot count.

Start the launcher for the model family you want to serve:

# Gemma 4 (default port 8082)
./scripts/serve_gemma4_opencode.sh

# Qwen 3.6 (default port 8081)
./scripts/serve_qwen36_opencode.sh

# DeepSeek-V4 (default port 8080)
./scripts/serve_deepseek4_opencode.sh

# A different explicitly supported GGUF can be served without hf2q provenance:
MODEL=./out/DeepSeek-V4-Flash-0731.gguf PORT=8090 \
  ./scripts/serve_deepseek4_opencode.sh

curl http://127.0.0.1:8080/v1/models

Point the client's OpenAI-compatible base URL at the selected launcher's http://127.0.0.1:<port>/v1 endpoint and select the model ID returned by /v1/models (normally the GGUF file stem). Set MAX_SLOTS=1 for one agent or MAX_SLOTS=8 for an eight-slot hardware experiment; four is the release-validated default. DeepSeek's CONTEXT_LEN override changes the full logical context of each slot; Gemma and Qwen use the context declared by their GGUF. KV_CACHE_BUDGET_BYTES independently caps aggregate physical KV high-water. Requests that cannot safely fit wait or fail explicitly instead of silently receiving a shorter context.

On the target M5 Max host, the launcher defaults to the schema-v2, source-bound deepseek4-agentic-q2 reproduction that passed the strict coherence, throughput, tool-use, and long-prefix cache gates. It enables operator progress telemetry and rejects unsafe port or memory state before mapping the approximately 100 GiB model.

Unary and streaming chat completions support reasoning content, OpenAI tools, required/automatic tool choice, parallel DSML invokes, cancellation, and usage telemetry. Growing transcripts reuse the live native KV/recurrent prefix; DeepSeek's old-reasoning canonicalization restores a prompt-tail checkpoint, so normal agent turns do not prefill the full context again. DeepSeek serving is slot-aware and uses bounded admission/decode waves so several agents make progress without duplicating model weights. Embeddings and multimodal messages remain unsupported for DeepSeek and fail explicitly rather than selecting another family or runtime.

For MoE models, pass an APEX tier instead of a standard ftype:

hf2q convert ./models/Qwen3.5-35B-A3B \
  --quant apex-balanced \
  -o ./out/qwen35-apex-balanced.gguf

The driver looks up the fingerprint manifest and, on match, logs [hf2q apex] auto-detected APEX config: vendor/apex-quant/configs/<file> before quantizing — confirming the exact per-tensor overlay in use.

I-tier APEX (imatrix-aware quantization)

The apex-i-* tiers (apex-i-quality, apex-i-balanced, apex-i-compact) require per-row activation-importance data (imatrix). Two ways to supply it:

# A. In-tree: hf2q runs its own forward pass over a calibration corpus.
#    Stage 3.0 supports Gemma 4 only; other arches use option B.
hf2q convert ./models/google-gemma-4-26b-a4b-it \
  --quant apex-i-balanced \
  --imatrix-corpus cdv3 \
  -o ./out/gemma4-26b-apex-i-balanced.gguf

# B. Pre-computed: pass an external `.imatrix.gguf` (works for any
#    supported arch — Qwen 3.5/3.6 MoE included).
llama-imatrix -m ./out/qwen35-f16.gguf \
  -f data/calibration/cdv3.txt \
  -o /tmp/qwen35.imatrix.gguf
hf2q convert ./models/Qwen3.5-35B-A3B \
  --quant apex-i-balanced \
  --imatrix /tmp/qwen35.imatrix.gguf \
  -o ./out/qwen35-apex-i-balanced.gguf

The in-tree path (option A) writes a temporary F16 GGUF, drives the forward pass over cdv3 (bartowski's calibration corpus, baked into the binary), and consumes the resulting per-tensor sum-of-squared-activations to choose the per-layer mix. Wall time is dominated by the forward pass: roughly seconds per 512-token chunk × ~100 chunks on a 26B-A4B Gemma 4 model = operator-coffee-time, not CI-time.

Optional flags:

• --imatrix-out <path> — write the computed (or loaded) imatrix to disk for reuse across multiple --quant apex-i-* runs against the same base model.
• --imatrix-n-ctx <N> — override the default 512-token chunk size (matches stock llama-imatrix -c 512). Larger N means fewer, longer chunks per forward-pass loop; useful when matching imatrices produced by stock llama-imatrix -c <other>. Must be > 0; passing 0 surfaces a typed ConvertError::ImatrixNCtxInvalid.

Architecture

A full source-grounded architecture map lives in docs/ARCHITECTURE.md. One-paragraph version:

   ┌──────────────┐    ┌──────────────────┐    ┌──────────────┐
HF │ input/       │ -> │ models/<arch>/   │ -> │ backends/    │
   │ - safetensors│    │ - tensor rename  │    │ - gguf       │
   │ - config     │    │ - MoE merge      │    │ - safetensors│
   └──────────────┘    │ - DWQ targets    │    └──────────────┘
                       └──────────────────┘            │ GGUF
                                                       v
                                              ┌──────────────────┐
                                              │ inference/       │
                                              │ - load + warmup  │
                                              │ - forward (mlx)  │
                                              │ - KV cache (TQ)  │
                                              │ - spec-decode    │
                                              └──────────────────┘
                                                       │
                                              ┌──────────────────┐
                                              │ serve/           │
                                              │ - OpenAI HTTP    │
                                              │ - SSE streaming  │
                                              │ - block-prefix$  │
                                              │ - multi-model    │
                                              └──────────────────┘

Performance

Re-bench at HEAD 2026-05-17 on M5 Max against llama.cpp peer (build 389ff61d7, -fa 1) with identical GGUFs. 3-run median; hf2q uses default config including the HF2Q_NO_FA hybrid-attn fix from commit 03328ee5. See docs/peer-parity-baselines-2026-04-26.md for the full thermal-fair alt-pair protocol used by ADR-029 baselines.

• Decode (Gemma-4 26B-A4B Q6_K) — tg200 1.01× peer-FA AHEAD (hf2q 105.2 t/s vs llama-bench 104.32 t/s); tg2000 0.97× peer-FA (hf2q 93.5 t/s vs 96.69 t/s). The historical ADR-029 iter-175 ~1.05× AHEAD across tg200/tg2000/tg5000 claim was measured at a pre-HF2Q_NO_FA HEAD; re-bench at current main shows it holding at tg200 only.
• Prefill (Gemma-4 26B) — crossover regime: pp1800 0.96× peer-FA (hf2q 2734 t/s vs llama-bench 2837 t/s); pp3700 1.24× peer-FA AHEAD (hf2q 2703 t/s vs 2181 t/s). hf2q's prefill rate drops ~1% from pp1800→pp3700 while llama's drops ~23% (FA tile-skip helps less at longer K), so the cross-over sits early in this range. The historical 1.07-1.09× AHEAD claim across the whole range no longer holds at current main.
• Decode (Qwen 3.6 35B-A3B APEX-Q5_K_M) — tg200 1.29× peer-FA AHEAD (hf2q 130.6 t/s vs 101.31 t/s). Historical ADR-028 ~1.34× measurement is within ~4% of current re-bench (thermal / build drift).
• KV-cache footprint — TurboQuant 8-bit (ADR-007 + ADR-027 iter-34) drops F32 K/V allocations entirely on Qwen 3.6 35B-A3B at 32K context, 340 MiB vs 1.34 GiB F32-only baseline = 3.94× memory savings. This is the only major performance claim with an in-tree regression pin (tests/qh35_no_f32_kv_alloc_with_tq_kv.rs).

Regression protection for the decode path: 8 parity tests (V2/V3 unbatched + V3 batched), coherence_smoke (2 cells), 200-token byte-identity verification. No automated bench-vs-peer gate is currently in CI — these numbers are operator-driven re-bench, not continuously verified.

Note: DWQ at the production-default perturb=1.0 is mathematically equivalent to the underlying K-quant baseline (ADR-020 finding 2026-05-08); DWQ wins materialize only at lower perturb values that move the scales/biases off the K-quant projection.

Performance work is investigation-driven and tracked in numbered ADR-029 (Gemma 4 decode), ADR-028 (peer-parity baseline), ADR-030 (speculative decode) iter-logs under docs/.

Repository layout

src/
├── arch/          single source of truth for per-arch conformance
├── backends/      GGUF + mlx-lm safetensors writers
├── calibrate/     DWQ training, autograd, imatrix
├── inference/     per-arch forward graphs, spec-decode, vision
├── input/         HF config + safetensors loaders, HF Hub download
├── intelligence/  hardware probe, auto-quant heuristics, RuVector
├── ir/            internal tensor / metadata representation
├── models/        per-arch tensor rename + MoE merge
├── quality/       cosine / KL / perplexity scorers
├── quantize/      Q-format codecs (legacy / K-quant / DWQ / mixed)
└── serve/         OpenAI HTTP API, block-prefix KV cache, multi-model
docs/              ADRs 004–031 + per-feature runbooks
tests/             77 integration test files
scripts/           109 bench / repro / runbook scripts

Development

cargo build              # debug build
cargo test               # full test suite
cargo build --release    # release binary
cargo run -- doctor      # diagnostic

The project is TDD-heavy: every ADR closes only when its acceptance tests + smoke prompts pass. New architectures must be onboarded via the checklist in docs/arch-onboarding.md — registry entry + tensor catalog + smoke prompt before any forward-pass code lands.

Documentation index

• docs/ARCHITECTURE.md — source-grounded architecture map.
• docs/converting-a-model.md — generic convert reference.
• docs/converting-qwen35.md — Qwen 3.5/3.6 specifics.
• docs/operating-kv-cache.md — TurboQuant KV cache operator guide.
• docs/operator-env-vars.md — every HF2Q_* env var, what it gates.
• docs/ADR-004…ADR-031 — every architectural decision, with rationale and verification status.

License

Dual-licensed under Apache-2.0 OR MIT (Cargo.toml license field; LICENSE-APACHE and LICENSE-MIT files at repo root). See docs/ADR-008-candle-divorce.md for the dependency philosophy.