An unofficial continuation of Saganaki22/ComfyUI-Hyperflow (Apache-2.0), whose GitHub repository is no longer available. Changes from the original: per-block curve fits for pruned bases and hybrids, the Create HyperFlow H3 Curve Fit node, and renamed node IDs (ApplyHyperFlowH3, ApplyHyperFlowH3Advanced, HyperFlowH3CurveFit) so it installs alongside the original. Workflows made with the original nodes need those nodes swapped.

中文文档 (README_ZH)

ComfyUI node pack for HyperFlow — Video Rebirth's 8-step LoRA for MiniMax-H3 (video + audio), ported onto ComfyUI's native MiniMax-H3 model. No ComfyUI core files are modified; no custom sampler is needed.

Note — don't want this custom node? Use the standalone LoRA builds instead — extracted and converted to plain ComfyUI format, they load with the stock Load LoRA node, nothing to install: drbaph/MiniMax-H3-Turbo-Lora-ComfyUI — minimax_h3_hyperflow_8step_v1.0_comfyui_bf16.safetensors (3.93 GB) / ..._pruned_bf16.safetensors (3.91 GB), plus rank-20 resized variants (~318/316 MB). Those apply the backbone LoRA only — the two-time (t, r) conditioning is exclusive to this node, so without it the output deviates from the released model.

The adapter is two things on top of the official base: a LoRA (rank 256, unmerged bf16 branches) and two-time (t, r) conditioning — every step is conditioned on the interval it integrates, r = 1 - sigma_next. This pack blends the two embeddings and preserves distinct (t, r) rows through the native blocks and output heads, including the first step where video and audio share t = 0. All forward patches and LoRA injections use ComfyUI's ModelPatcher; no core files are changed.

https://github.com/user-attachments/assets/3e2e2d5c-6a39-46e9-8946-aa42f08c3d9d

Install

Clone into ComfyUI/custom_nodes/ComfyUI-HyperFlow-H3.
Download one converted weights file from the drbaph/Hyperflow-Comfyui Hugging Face repo (or just enable the node's download_if_missing toggle and it fetches itself on first run):

📂 ComfyUI/
└── 📂 models/
    └── 📂 hyperflow/
        ├── custom_node_hyperflow_8step_v1.0_comfyui.safetensors         (3.67 GiB, full base — the released 8-step model)
        └── custom_node_hyperflow_8step_v1.0_comfyui_pruned.safetensors  (3.64 GiB, pruned/curve bases — backbone only, single-time)

The hyperflow.json manifests ship with this pack in assets/ — users only ever download the .safetensors. Converted files carry the full self-describing HyperFlow header (gate, sigma grid, rank); the original diffusers-layout file is rejected with a clear message, never translated at load time.

Restart ComfyUI.

Supported models

Full MiniMax-H3 models (the regular, unpruned downloads such as minimax_h3_fl2va_int8_convrot or minimax_h3_ref2va_int8_convrot): fully supported. Use the full adapter, custom_node_hyperflow_8step_v1.0_comfyui.safetensors, and you get HyperFlow exactly as released.

Pruned MiniMax-H3 models (smaller files with part of the model removed to save memory): use the _pruned adapter. Pruning removes the part of the model HyperFlow uses to keep track of each step, so on its own HyperFlow can only apply its 8-step LoRA there. For the models below, turn on experimental_curve_refit and a small curve fit bundled with this pack puts most of that back; there is nothing extra to download:

minimax_h3_fl2va_pruned_int8_convrot
minimax_h3_ref2va_pruned_int8_convrot
The FL2VA/REF2VA hybrid merges minimax_h3_hybrid_fl2va_ref2va_b15-49-int8, _b20-49-int8, _b25-49-int8 and _b30-49-int8
Minimax-h3_Singularity_ref2va_Pruned_v1.3_int8

A curve fit only matches the exact file it was made from. Renaming your copy is fine, but a re-uploaded or re-quantized version of the same model is a different file: the console then says curve refit disabled and HyperFlow runs LoRA-only.

Other pruned models (merges and finetunes) work LoRA-only. If the model's author also publishes the unpruned version, you can make your own curve fit with the Create HyperFlow H3 Curve Fit node; see CURVE_FITS.md.

Usage

Load Diffusion Model (MiniMax-H3)
  -> ApplyHyperFlowH3            (MODEL -> MODEL + SIGMAS)
  -> [optional] Model Attention Backend        (core node; dense fallback incl. comfy-kitchen int8)
  -> [optional] Model Sparse Attention         (core node; sol-attn / sla / vsa)
  -> SamplerCustomAdvanced + guider + Euler    (feed the SIGMAS output, not a scheduler)

ApplyHyperFlowH3 outputs the trained 9-point sigma grid as SIGMAS — wire it into SamplerCustomAdvanced in place of BasicScheduler. Use Euler for the released recipe; other samplers are ablations. The two-time endpoints are derived from sample_sigmas.
lora_mode: bypass (default) applies the LoRA at run time — matches the reference's unmerged bf16 branches. merge folds the backbone LoRA into the weights; the two small base time projections stay in bypass so the endpoint branch reads unmodified weights. Merging quantized weights can change numerical results.
download_if_missing fetches the chosen variant (auto matches the detected base) from drbaph/Hyperflow-Comfyui into models/hyperflow/ — exactly the published .safetensors, nothing else. Off by default.
ApplyHyperFlowH3Advanced adds gate and sigma-grid overrides (ablations; defaults reproduce the released model).

Sol-Attn (optional sparse attention) — core node settings

HyperFlow's validated Sol-Attn recipe maps onto the core Model Sparse Attention node:

| HyperFlow (reference) | Model Sparse Attention | | --- | --- | | dense_steps = 2 (of 8) | start_percent = 0.16 | | dense_layers = (0, 1) | dense_blocks = "0,1" | | tau = 1.0 | tau = 1.0 (method sol-attn) | | sink tokens: none | sink_conditioning = "off" | | — | extra_tokens = 0 |

start_percent is measured on the unshifted noise clock, not as a fraction of steps. With the shipped grid and shifts, 0.16 keeps exactly the first two evaluations dense. Sparse attention itself is an upstream-supported approximation; use dense attention for the dense recipe.

SLA is a different sparse method — use it only with SLA-trained weights.

Pruned / curve bases — experimental curve refit

Pruned MiniMax-H3 bases have no time_embedder, so the two-time (t, r) pathway cannot apply — by default the pruned build applies the backbone LoRA only and runs single-time (off-recipe). The experimental curve refit restores most of the two-time signal through a small checkpoint-bound fit (~1.7 MB, one set of values per block; no download, no extra model weights). Fits are bundled in assets/curve_fits/ for the FL2VA and REF2VA pruned int8 checkpoints, the FL2VA/REF2VA block-spliced hybrids (b15/b20/b25/b30-49) and Singularity REF2VA pruned v1.3 — see the list.

https://github.com/user-attachments/assets/e72c45e3-2cc3-48ce-bb1a-2f0c25e7d602

Recommended settings for pruned:

Enable experimental_curve_refit (widget on both Apply nodes; right-click → show optional widgets if hidden), keep strength = 1.0, and leave gate/sigma overrides empty. Matching is tiered: the exact fitted files apply silently; byte-different copies of the fitted base or adapter (mirrors, HF downloads) still apply best-effort with a console warning — the recipe (strength/gate/sigmas) is always validated; unknown checkpoints or a modified MODEL fall back to backbone-only, with the file hashes in the log for support.
Use bypass as lora_mode for the sharpest match to the reference branches.
Same sampling as full bases: Euler + the node's SIGMAS output + the Sol-Attn table above.
Console: [hyperflow] curve refit disabled: <reason> means the fallback engaged — check strength/overrides/checkpoint.

Measured against the full base with the earlier shared (one vector for all blocks) fits (fixed prompt/seed/latents): FL2VA video cosine 0.81 → 0.88, REF2VA audio cosine 0.93 → 0.99. It recovers most — not all — of the missing conditioning; the full base remains the reference. Details and per-family numbers: validation/curve_findings.md. Other pruned checkpoints, including block-spliced hybrids, need their own fit: generate one with the Create HyperFlow H3 Curve Fit (Pruned Bases) node (or tools/fit_curve.py), which writes it to models/hyperflow/curve_fits/ — see CURVE_FITS.md.

Notes

Existing node weights are corrected on load: the loader swaps legacy Diffusers SwiGLU rows into ComfyUI order. No replacement download is needed. New conversions carry hyperflow_fc1_layout=gate_value and are not swapped again.
Bypass QKV adapters store the three separate LoRA projections instead of the block-diagonal zero padding, saving 1,092 MiB of adapter tensor storage for the released rank-256 build.
Base detection is automatic: the node inspects the loaded model — full base (has time_embedder) or pruned/curve base (no time_embedder) — and enforces the matching weights build with a clear error that names the right file. The pruned-base build applies the backbone LoRA only and runs single-time (off-recipe).
Experimental curve refit for pruned/curve bases: see the dedicated section above.
Quantized bases (int8/fused ops): LoRA targets that the base folds into a fused kernel (no hookable module) are detected and applied through the merge path automatically — the console report lists them as N fused/int8 targets via merge.
Model sampling shifts: the H3 model already defaults to video shift 12 / audio shift 3. The core ModelSampling node goes after ApplyHyperFlowH3 in the chain, and is only needed if you want non-default shifts.
aimdo malloc-graph: on Comfy builds whose model compiler crashes on patched MiniMax-H3 forwards, the node disables the compiler for its own model calls.
Weights are a Model Derivative of MiniMax-H3 under the MiniMax H3 Community License; this pack's code is Apache-2.0 (the schedule/embedder ports derive from the HyperFlow and diffusers code, see the upstream THIRD_PARTY_NOTICES.md).

Stuck at “Model Initializing” after a ComfyUI update

That status covers the whole first sampling step, including deferred weight loading and LoRA merging. It does not identify a compiler hang. merge reduces resident adapter memory, but still needs temporary memory and time to patch the base weights.

ComfyUI v0.37 introduced automatic fast-disk detection compared with v0.36. To isolate a stall after updating, restart ComfyUI with --disable-fast-disk and retry the same workflow. Treat this as a diagnostic, not a confirmed HyperFlow fix. If it still stalls, report the ComfyUI commit, full console log, base/LoRA filenames, GPU/VRAM and system RAM, and whether it happens on the first chunk or a later one. Include the [hyperflow] ... applied line if present.

Acknowledgements

MiniMax-H3 (GitHub): the base model, VAEs, conditioner and official workflows.
AnyFlow (GitHub; Gu et al., 2026): the flow-map formulation behind the two-time (t, r) conditioning.
Sol-Attn (GitHub; Li et al., 2026): optional attention kernels.

Thanks to their authors.