videorebirth commited on
Commit
43feb5e
·
0 Parent(s):

Release minimax_h3_hyperflow_8step_v1.0.safetensors

Browse files
Files changed (6) hide show
  1. .gitattributes +35 -0
  2. LICENSE +84 -0
  3. NOTICE +13 -0
  4. README.md +243 -0
  5. hyperflow.json +9 -0
  6. minimax_h3_hyperflow_8step_v1.0.safetensors +3 -0
.gitattributes ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
LICENSE ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MiniMax H3 COMMUNITY LICENSE AGREEMENT
2
+ MiniMax H3 release date/License date: August 2, 2026.
3
+ The scope of this License Agreement (this “Agreement”) is expressly limited to the “Applicable Territory” as defined below.
4
+ By clicking to accept, or by using, reproducing, modifying, distributing, running, or displaying any portion or element of the MiniMax H3 Works (including through any Hosted Services) in any manner, you acknowledge and accept the terms of this Agreement, and this Agreement shall take immediate effect upon the occurrence of such act.
5
+ I. Definitions
6
+ 1. “Acceptable Use Policy” means the policy published by MiniMax in Exhibit A.
7
+ 2. “Agreement” means the terms and conditions set forth herein that govern the use, reproduction, distribution, modification, running, and display of the MiniMax H3 Works or any portion or element thereof.
8
+ 3. “Applicable Territory” means worldwide, excluding the Excluded Territories.
9
+ 4. “Documentation” means the specifications, manuals, and documentation concerning MiniMax H3 that are publicly released by MiniMax.
10
+ 5. “Excluded Territories” means the European Union, the United Kingdom, the Republic of Korea and the United States of America.
11
+ 6. “MiniMax H3” means the video generation model, together with its software and algorithms, including trained model weights, parameters (including optimizer states), machine-learning model code, inference-supporting code, and other elements thereof made publicly available by Us, as released at https://huggingface.co/MiniMaxAI/MiniMax-H3.
12
+ 7. “MiniMax H3 Works” means (i) the Materials, (ii) the Model Derivatives, and (iii) all derivatives thereof.
13
+ 8. “Hosted Services” means hosted services provided via application programming interfaces (APIs), web access, or any other electronic or remote means.
14
+ 9. “Licensee,” “you,” or “your” means the natural or legal person exercising rights and/or using the MiniMax H3 Works for any purpose in any field of use under this Agreement.
15
+ 10. “Materials” means, collectively, MiniMax H3 and the Documentation (and any portion thereof), in each case as made available by MiniMax under this Agreement and proprietary to MiniMax.
16
+ 11. “Model Derivatives” means all of the following: (i) any modification of MiniMax H3 or any Model Derivative thereof; (ii) any work based on MiniMax H3 or any Model Derivative thereof; or (iii) any other machine learning model created by transferring the patterns of the weights, parameters, operational patterns, or Outputs of MiniMax H3 or any Model Derivative thereof to another model, such that the latter model exhibits behavior similar to MiniMax H3 or its Model Derivatives, including by distillation methods, methods using intermediate data representations, or methods based on training using synthetic-data Outputs generated by MiniMax H3 or its Model Derivatives. For the avoidance of doubt, Outputs are not deemed Model Derivatives.
17
+ 12. “Output” means any result of operating or otherwise using MiniMax H3 or any Model Derivatives (including through Hosted Services).
18
+ 13. “Third Party” means any natural or legal person that is not under common control with us or with you.
19
+ 14. “Including” means “including but not limited to.”
20
+ 15. “We,” “Us” or “MiniMax” means Nanonoble Pte. Ltd..
21
+ II. Grant of Rights
22
+ Solely within the Applicable Territory, we grant you a non-exclusive, non-transferable, royalty-free, limited license to use, reproduce, distribute, create derivative works (including Model Derivatives), and modify the Materials in accordance with the terms of this Agreement and the Acceptable Use Policy, based on the intellectual property and other rights owned by MiniMax that are embodied in or used by the Materials. You shall not violate (or encourage or permit any person to violate) any term of this Agreement or the Acceptable Use Policy.
23
+ We will continuously evaluate the applicable laws, regulations and compliance requirements for the Excluded Territories. In the meantime, should any person in such Excluded Territories be interested in deploying our models, you are welcome to contact us about obtaining a license, which will be granted based on robust controls and guardrails for purposes of complying with the laws, regulations and compliance requirements of the Excluded Territories.
24
+ III. Distribution and Redistribution
25
+ Subject to and conditioned on your continuing compliance with this Agreement, including its territorial restrictions and the Acceptable Use Policy, and solely within the Applicable Territory, you may distribute or make available the MiniMax H3 Works to Third Parties within the Applicable Territory; provided, that all of the following conditions are met:
26
+ 1. You must provide a copy of this Agreement to all such Third Parties who receive the MiniMax H3 Works or use your products or services related thereto;
27
+ 2. You must cause any modified files to carry prominent notices stating that you have modified such files;
28
+ 3. You are encouraged to:
29
+ a. display a notice on any product or service developed using MiniMax H3 indicating that the product or service is “Powered by MiniMax H3”;
30
+ b. add an AI-generation identifier to files produced using generative AI models including MiniMax H3; and
31
+ c. publish at least one technical blog post or a public statement describing your experience using MiniMax H3 Works;
32
+ 4. All distributions to Third Parties (other than through Hosted Services) must be accompanied by a “NOTICE” text file containing the following notice:
33
+ “MiniMax H3 is licensed under the MiniMax H3 Community License Agreement, Copyright © 2026 MiniMax. All Rights Reserved.”
34
+ You may add your own copyright notices on your modifications; except as provided in this Section and in Section V, however, you may not impose additional or different terms and conditions on the use, reproduction, or distribution of your modifications or of any aggregate Model Derivatives, and your use, reproduction, modification, distribution, running, and display of the work must otherwise comply with the terms and conditions of this Agreement (including the provisions concerning the Applicable Territory). If you receive the MiniMax H3 Works from a Licensee as part of an integrated end-user product, the provisions of Section III of this Agreement do not apply to you, but Section V and Exhibit A remain applicable.
35
+ IV. Additional Commercial Terms
36
+ 1. You shall obtain a separate, prior written authorization from MiniMax by contacting api@minimax.io with the subject line “MiniMax H3 licensing - authorization request”, if your commercial products and services generate more than 20 million US dollars (or equivalent in other currencies) in yearly revenue.
37
+ 2. You shall prominently display “MiniMax H3”on the user interface of commercial product or service that uses MiniMax H3 or MiniMax H3 Works.
38
+ V. Use Restrictions
39
+ 1. Your use of the MiniMax H3 Works must comply with applicable laws and regulations (including trade-compliance laws and regulations) and must comply with the Acceptable Use Policy for the MiniMax H3 Works, which is incorporated into this Agreement by reference.
40
+ 2. Before providing access to the MiniMax H3 Works or any product, service, or Hosted Service incorporating them, you must bind each recipient or user to enforceable terms at least as protective as the use restrictions in this Section V and Exhibit A, and you must notify each recipient or user that those restrictions apply.
41
+ 3. You may not use the MiniMax H3 Works or any of their Outputs or results to improve any other artificial intelligence model (other than MiniMax H3 or its Model Derivatives).
42
+ 4. You may not use, reproduce, modify, distribute, or display the MiniMax H3 Works or any of their Outputs or results outside the Applicable Territory. Any such use outside the Applicable Territory is not authorized by this Agreement.
43
+ 5. If you provide or make available to any Third Party a product, service, or Hosted Service that permits the generation of Outputs using MiniMax H3 or any Model Derivative, you must, before making that product or service available and throughout its operation, implement, maintain, test, and periodically review reasonable and proportionate technical and organizational safeguards designed to prevent and mitigate access, uses, and Outputs that violate this Section V or Exhibit A, including uses or Outputs that infringe, misappropriate, or otherwise violate any Third Party’s intellectual-property or other rights. You must not knowingly disable, materially weaken, or permit the circumvention of those safeguards. You must maintain a reasonably accessible mechanism for reporting suspected violations. Upon receiving a good-faith report or otherwise obtaining actual knowledge of a violation, you must promptly investigate and take reasonable steps within your control to stop or mitigate the violation, including removing or disabling access to offending content or services and suspending or terminating repeat violators where appropriate. You are responsible for implementing and enforcing these requirements with respect to your products, services, systems, users, and downstream recipients.
44
+ VI. Intellectual Property
45
+ 1. Subject to MiniMax’s rights in the MiniMax H3 Works (and the intellectual property therein), and to your compliance with the terms and conditions of this Agreement, as between you and MiniMax, you will own the derivative works and modifications of the Materials that you have created or had created, as well as any Model Derivatives.
46
+ 2. Except for the limited license expressly granted in this paragraph, no trademark license is granted under this Agreement; with respect to MiniMax H3 Works, the Licensee may not use any name or mark owned by or associated with MiniMax or any of its affiliates, except as reasonably and customarily necessary to describe and distribute the MiniMax H3 Works. MiniMax hereby grants you a license to use the “MiniMax H3” mark (the “Mark”) within the Applicable Territory solely for the purpose of complying with Section III.3; provided, that you comply with all applicable trademark-protection laws. All goodwill arising from your use of the Mark shall inure to the benefit of MiniMax.
47
+ 3. If you bring or assert any suit or other legal proceeding (including a cross-claim or counterclaim in any action) against us or any other natural or legal person alleging that the Materials, any Output, or any portion of the foregoing infringes any intellectual property right or other right owned by you or for which you can obtain a license, all licenses granted to you under this Agreement will terminate as of the date such suit or proceeding is filed. You shall defend, indemnify, and hold us harmless against any Third-Party claim arising out of or related to the use or distribution of the MiniMax H3 Works by you or by any Third Party.
48
+ 4. MiniMax claims no rights over the Outputs you generate. You and your users are entirely responsible for the Outputs and any subsequent use thereof.
49
+ VII. Disclaimers and Limitations of Liability
50
+ 1. We have no obligation to support, update, provide training for, or develop any further version of the MiniMax H3 Works, or to grant any license with respect thereto.
51
+ 2. UNLESS AND ONLY TO THE EXTENT REQUIRED BY APPLICABLE LAW, THE MINIMAX H3 WORKS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED “AS IS” WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES OF ANY KIND INCLUDING ANY WARRANTIES OF TITLE, MERCHANTABILITY, NONINFRINGEMENT, COURSE OF DEALING, USAGE OF TRADE, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING, REPRODUCING, MODIFYING, PERFORMING, DISPLAYING OR DISTRIBUTING ANY OF THE MINIMAX H3 WORKS OR OUTPUTS AND ASSUME ANY AND ALL RISKS ASSOCIATED WITH YOUR OR A THIRD PARTY’S USE OR DISTRIBUTION OF ANY OF THE MINIMAX H3 WORKS OR OUTPUTS AND YOUR EXERCISE OF RIGHTS AND PERMISSIONS UNDER THIS AGREEMENT.
52
+ 3. TO THE FULLEST EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT SHALL MINIMAX OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, FOR ANY DAMAGES, INCLUDING ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, EXEMPLARY, CONSEQUENTIAL OR PUNITIVE DAMAGES, OR LOST PROFITS OF ANY KIND ARISING FROM THIS AGREEMENT OR RELATED TO ANY OF THE MINIMAX H3 WORKS OR OUTPUTS, EVEN IF MINIMAX OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.
53
+ VIII. Term and Termination
54
+ 1. This Agreement is effective from the moment you accept this Agreement or begin accessing the Materials, and, subject to your compliance with its terms and conditions, will remain in effect until terminated as provided herein.
55
+ 2. If you breach any term or condition of this Agreement, we have the right to terminate this Agreement. Upon termination, you must immediately cease accessing, using, and distributing the MiniMax H3 Works; delete or destroy all copies within your possession or control; and notify each downstream recipient that your authorization has ended. The obligations in the preceding sentence and Sections VI.1, VI.3, VII, and IX survive termination.
56
+ IX. Governing Law and Jurisdiction
57
+ 1. This Agreement, and any dispute arising out of or related to this Agreement, shall be governed by the laws of the Hong Kong Special Administrative Region of the People’s Republic of China, without regard to its conflict-of-laws rules. The United Nations Convention on Contracts for the International Sale of Goods does not apply to this Agreement.
58
+ 2. Any dispute arising out of or related to this Agreement shall be subject to the exclusive jurisdiction of the courts of the Hong Kong Special Administrative Region of the People’s Republic of China with competent jurisdiction. Both MiniMax and the Licensee hereby consent to the exclusive jurisdiction of such courts for any such dispute.
59
+ Additional Note: Please note that the encoder of MiniMax H3 uses Qwen3-VL-32B, which is licensed under Apache 2.0 License: https://github.com/QwenLM/Qwen3-VL/blob/main/LICENSE.
60
+
61
+ Exhibit A — Acceptable Use Policy
62
+ MiniMax reserves the right to update this Acceptable Use Policy from time to time.
63
+ Last revised: August 2, 2026.
64
+ MiniMax is committed to promoting the safe and fair use of its tools and features, including MiniMax H3. You agree not to use MiniMax H3, any Model Derivatives, or any Output in any of the following ways:
65
+ 1. Use outside the Applicable Territory;
66
+ 2. Use in any manner that violates any applicable national, federal, state, local, or international law, regulation, or other legal requirement, or that infringes, misappropriates, or otherwise violates any Third Party’s intellectual-property or other proprietary rights, including through unauthorized reproduction, distribution, public display, public performance, or creation of derivative works;
67
+ 3. Use in any manner that may harm yourself or others;
68
+ 4. Use to repurpose or distribute the Outputs of MiniMax H3 or any Model Derivatives in order to harm yourself or others;
69
+ 5. Use to circumvent or bypass any safety guardrails or safeguards we have implemented;
70
+ 6. Use in any manner that exploits or harms, or intends to exploit or harm, minors;
71
+ 7. Use to generate or disseminate verifiably false information and/or content for the purpose of harming others or influencing elections;
72
+ 8. Use to manufacture or facilitate false online engagement, including fake reviews and other means of false online engagement;
73
+ 9. Use to intentionally defame, disparage, or otherwise harass others;
74
+ 10. Use to generate and/or disseminate malware (including ransomware) or any other content intended to damage electronic systems;
75
+ 11. Use to generate or disseminate personally identifiable information for the purpose of harming others;
76
+ 12. Use to generate or disseminate information (including images, code, posts, or articles) in or to any public environment (including via bot tweets or similar means) without clearly and prominently disclosing that such information and/or content is machine-generated;
77
+ 13. Use to impersonate another person without that person’s consent, authorization, or lawful right to do so;
78
+ 14. Use to make high-risk automated decisions in critical domains that affect individual safety, rights, or well-being (such as law enforcement, immigration, healthcare or medical services, critical-infrastructure management, product-safety components, essential services, credit, employment, housing, education, social scoring, or insurance);
79
+ 15. Use in any manner that violates or disregards the social, ethical, or moral standards of other countries or regions;
80
+ 16. Use to carry out, assist, threaten, incite, plan, advocate for, or encourage violent extremism or terrorism;
81
+ 17. Use for any purpose intended to discriminate against, or harm, individuals or groups based on protected characteristics or categories, online or offline social behavior, or known or predicted personality traits;
82
+ 18. Use to intentionally exploit the vulnerabilities of specific populations based on age, social, physical, or psychological characteristics, so as to materially distort the behavior of a member of that group in a manner that causes, or is likely to cause, physical or psychological harm to that person or to others;
83
+ 19. Use for military purposes;
84
+ 20. Use to engage in any unauthorized or unlicensed professional activity, including but not limited to financial, legal, medical or healthcare, or other professional practice.
NOTICE ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MiniMax H3 is licensed under the MiniMax H3 Community License Agreement, Copyright © 2026 MiniMax. All Rights Reserved.
2
+
3
+ The HyperFlow LoRA weights in this repository (minimax_h3_hyperflow_*.safetensors) are a Model Derivative of
4
+ MiniMax-H3 (https://huggingface.co/MiniMaxAI/MiniMax-H3) as defined in Section I.11 of that Agreement. They were
5
+ created by the HyperFlow authors (Copyright 2026 The HyperFlow authors) and are distributed under the same
6
+ MiniMax H3 Community License Agreement, a copy of which is included in this repository as LICENSE.
7
+
8
+ Modifications (Section III.2): each LoRA file is a new file trained against MiniMaxAI/MiniMax-H3 at the revision
9
+ recorded in its header (`base_model_revision`). No file of MiniMax-H3 is modified or redistributed here; the base
10
+ weights, VAEs and text encoder remain the official ones.
11
+
12
+ HyperFlow is not an official MiniMax product and is not affiliated with or endorsed by MiniMax.
13
+ Powered by MiniMax H3.
README.md ADDED
@@ -0,0 +1,243 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ license_name: minimax-h3-community-license-agreement
4
+ license_link: LICENSE
5
+ base_model: MiniMaxAI/MiniMax-H3
6
+ base_model_relation: adapter
7
+ pipeline_tag: image-text-to-video
8
+ library_name: hyperflow-h3
9
+ tags:
10
+ - lora
11
+ - minimax-h3
12
+ - diffusers
13
+ - video-generation
14
+ - audio-video-generation
15
+ - text-to-video
16
+ - image-to-video
17
+ - image-text-to-video
18
+ - text-to-audio-video
19
+ - image-to-audio-video
20
+ - reference-to-audio-video
21
+ - flow-matching
22
+ - distillation
23
+ - self-distillation
24
+ - few-step
25
+ ---
26
+
27
+ # HyperFlow for MiniMax-H3
28
+
29
+ > By downloading or using these weights you accept the [MiniMax H3 Community License Agreement](LICENSE), including
30
+ > its territorial scope (not licensed in the EU, UK, South Korea or the US without MiniMax's separate authorization)
31
+ > and its Acceptable Use Policy. See [License](#license). Not an official MiniMax product.
32
+
33
+ HyperFlow is [Video Rebirth](https://www.videorebirth.com/)'s 8-step LoRA for
34
+ [MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3), obtained by data-free flow self-distillation and running
35
+ on the official [diffusers Modular Pipeline](https://huggingface.co/docs/diffusers/main/api/pipelines/minimax_h3).
36
+ Diffusers' default 50-point sigma schedule performs 49 model forwards; HyperFlow performs 8. The base weights, VAEs,
37
+ conditioner and workflows (`t2va`, `fl2va`, `ref2va`, all video + audio) stay the official ones. This repository
38
+ holds the LoRA only; the loader and examples live at
39
+ [github.com/Video-Rebirth/hyperflow](https://github.com/Video-Rebirth/hyperflow).
40
+
41
+ | | |
42
+ |---|---|
43
+ | Base model | [`MiniMaxAI/MiniMax-H3`](https://huggingface.co/MiniMaxAI/MiniMax-H3) (revision recorded in the file header) |
44
+ | Adapter | PEFT LoRA, rank 256 / alpha 256, on attention, feed-forward and both time embedders (316 modules); 2.8 GB |
45
+ | Sampling | 8 forward passes on a fixed sigma grid stored in the file (video shift 12, audio shift 3) |
46
+ | Workflows | `t2va`, `fl2va` and `ref2va`, one file for all three (`ref2va` loads it onto `transformer_ref/`) |
47
+ | Speed-up | ~3x end to end against the 49-NFE Diffusers baseline (~60 s vs ~175 s per clip on 4x H200) |
48
+ | Optional | Ulysses context parallel on up to 4 GPUs; FlashAttention-3; NVIDIA Sol-Attn sparse attention |
49
+ | License | [MiniMax H3 Community License Agreement](LICENSE); code on GitHub is Apache-2.0. See [License](#license) |
50
+
51
+ **Not a generic LoRA.** The file carries keys for a second time embedder that only exists once the HyperFlow loader
52
+ has installed it, and the step count comes from the file. Load it with `load_hyperflow_lora` from the
53
+ `hyperflow-h3` package, not with diffusers' `load_lora_weights` / `load_lora_adapter`.
54
+
55
+ ## Files
56
+
57
+ | File | `hyperflow_version` | Notes |
58
+ |---|---|---|
59
+ | `minimax_h3_hyperflow_8step_v1.0.safetensors` | `1.0` | **default**: what a bare repo id loads. First release |
60
+ | `hyperflow.json` | | Manifest: the `default` file and every file's `hyperflow_version` and `sha256` |
61
+ | `LICENSE`, `NOTICE` | | MiniMax H3 Community License Agreement and the notice it requires |
62
+
63
+ A published file name is never reused: pin one with `filename=` (or `--weights-filename`) and it stays what it is.
64
+ A bare repo id follows the manifest's `default`, so it picks up new versions as they ship; `revision=` pins a tag or
65
+ commit of this repository.
66
+
67
+ ## Usage
68
+
69
+ ```bash
70
+ pip install "hyperflow-h3[examples] @ git+https://github.com/Video-Rebirth/hyperflow.git"
71
+ ```
72
+
73
+ Python ≥ 3.10, diffusers ≥ 0.40.0 (the first release with the MiniMax-H3 Modular Pipeline and its context parallel
74
+ plan), transformers ≥ 4.57 (Qwen3-VL). The `examples` extra adds PyAV for writing the output video.
75
+
76
+ ```python
77
+ import torch
78
+ from diffusers import ComponentsManager
79
+ from diffusers.utils import load_image
80
+ from hyperflow_h3 import hyperflow_blocks, load_hyperflow_lora
81
+
82
+ manager = ComponentsManager()
83
+ blocks = hyperflow_blocks("fl2va") # official workflow, two blocks swapped
84
+ pipe = blocks.init_pipeline("MiniMaxAI/MiniMax-H3", components_manager=manager)
85
+ pipe.load_components(dtype=torch.bfloat16)
86
+ load_hyperflow_lora(pipe, "videorebirth/hyperflow") # LoRA + two-time embedder + 8-step grid
87
+
88
+ # The official single-GPU recipe (auto CPU offload), armed only now: registering a component resets the margin.
89
+ # 24 GB is the margin validated on H200; the official recipe's 12 GB starves the denoiser there. See Performance.
90
+ manager.enable_auto_cpu_offload(device="cuda", memory_reserve_margin="24GB")
91
+
92
+ out = pipe(
93
+ prompt="A red fox trotting through a snowy pine forest",
94
+ image=load_image("first.png"),
95
+ num_frames=124,
96
+ generator=torch.Generator().manual_seed(42),
97
+ output=["videos", "audio", "sampling_rate"],
98
+ )
99
+ ```
100
+
101
+ - Do not pass `num_inference_steps`: the step count and sigma grid come from the weights file; another count raises.
102
+ - Every other input (`prompt`, `image`, `last_image`, `references`, `height`, `width`, `num_frames`, `generator`)
103
+ is the official pipeline's.
104
+ - Other workflows: `hyperflow_blocks("t2va")` (no `image`) and `hyperflow_blocks("ref2va")` (`references=`; the
105
+ same weights file is loaded onto `transformer_ref`).
106
+ - Prompting: MiniMax's hosted H3-Context-IR prompt refinement is not part of the open weights, so prompts follow the
107
+ base model's [prompting guidance](https://huggingface.co/MiniMaxAI/MiniMax-H3#prompting-guidance) as they do in
108
+ the base Diffusers pipeline.
109
+
110
+ **Pinning and offline use.** `load_hyperflow_lora` takes this repo id, a local directory or a `.safetensors` file.
111
+ Manifest and weights are fetched once into the `huggingface_hub` cache and keep working under `HF_HUB_OFFLINE=1`.
112
+
113
+ ```python
114
+ load_hyperflow_lora(pipe, "videorebirth/hyperflow", filename="minimax_h3_hyperflow_8step_v1.0.safetensors")
115
+ load_hyperflow_lora(pipe, "/weights/hyperflow") # from `hf download videorebirth/hyperflow --local-dir /weights/hyperflow`
116
+ ```
117
+
118
+ **Example commands, several GPUs, Sol-Attn.** The package installs `hyperflow-h3-fl2va` and
119
+ `hyperflow-h3-ref2va`; the GitHub repository also has [`examples/`](https://github.com/Video-Rebirth/hyperflow/tree/main/examples)
120
+ source-tree launchers. They provide Ulysses sequence parallel over up to 4 GPUs (the degree MiniMax serves the model
121
+ with), FlashAttention-3, NVIDIA Sol-Attn and `--baseline` (Diffusers' default 50 sigma points / 49 NFE on the same
122
+ seed, for an A/B). Fully offline runs take a local weights file and a local MiniMax-H3 snapshot.
123
+
124
+ ```bash
125
+ hyperflow-h3-fl2va --prompt "..." --image first.png # first frame -> video+audio, up to 4 GPUs
126
+ hyperflow-h3-ref2va --prompt "..." --ref subject.png --ref clip.mp4 --ref voice.wav
127
+ hyperflow-h3-fl2va --prompt "..." --image first.png --gpus 1 # the official single-GPU recipe
128
+ ```
129
+
130
+ ## Performance
131
+
132
+ HyperFlow (8 NFE, 9 sigma points) against the base pipeline at Diffusers' default (49 NFE, 50 sigma points) on the
133
+ same clip: `fl2va`, 124 frames, 1344x768, seed 0, dense SDPA attention. Pipeline time covers text encoding, denoising
134
+ and decoding; process start-up and weight loading add ~30 s.
135
+
136
+ | | Base, 49 NFE | HyperFlow, 8 NFE | Speed-up |
137
+ |---|---|---|---|
138
+ | 4x H200, Ulysses degree 4 | ~175 s | ~60 s | 2.9x |
139
+ | 1x H200, auto CPU offload | ~395 s | ~130 s | 3.0x |
140
+
141
+ The speed-up is below the 49:8 NFE ratio because text encoding and VAE decoding, which HyperFlow leaves unchanged,
142
+ now take much of the run. Peak accelerator memory is ~80 GB per card either way (the LoRA adds 2.8 GB). The
143
+ single-GPU row uses `memory_reserve_margin="24GB"`, not the official recipe's 12 GB: on a 141 GB H200 a smaller margin
144
+ evicts the 10 GB VAE instead of the 62 GB text encoder when the DiT arrives and starves the denoiser. An 80 GB card
145
+ is untested.
146
+
147
+ With a fixed seed a run reproduces itself (audio bit-identical, video PSNR ≈ 43–50 dB, a residue of the VAE decode).
148
+ Across GPU counts or attention backends the clip is only visibly similar (PSNR ≈ 27 dB and ≈ 34 dB): sharding and
149
+ kernels change the bf16 reduction order. NVIDIA Sol-Attn sparse attention, optional and off by default, is a further
150
+ approximation (PSNR ≈ 23 dB against the dense run); check its output before adopting it.
151
+
152
+ ## Quality
153
+
154
+ HyperFlow is a self-distillation: the base model is the only teacher. Against the base pipeline at 49 NFE, on the
155
+ same seeds, four things stand out:
156
+
157
+ - **More balanced capabilities**
158
+ - **Better camera control**
159
+ - **Better consistency**
160
+ - **Better materials and detail**
161
+
162
+ <!-- TODO(comparison): side-by-side clips, base 49 NFE vs HyperFlow 8 NFE, same seed, 124 frames at 1344x768.
163
+ | Prompt | Base, 49 NFE | HyperFlow, 8 NFE |
164
+ |---|---|---|
165
+ | ... | ... | ... |
166
+ -->
167
+
168
+ Judge on your own prompts: `--baseline` in the example scripts renders both pipelines on one seed.
169
+
170
+ ## How it works
171
+
172
+ Three additions to the official pipeline, all reversible (`disable_hyperflow`):
173
+
174
+ - **Two-time conditioning.** The LoRA was trained with both the current time `t` and the step's endpoint `r` (a
175
+ flow-map formulation). A `TwoTimeEmbedder` wraps the base `time_embedder`, adds a LoRA'd copy for `r`, and blends
176
+ the two embeddings with a fixed gate stored in the file. Conditioning rows (`fl2va` keyframes, `ref2va`
177
+ references) keep `r = t`, so they stay pinned exactly as in the official denoiser.
178
+ - **8-step schedule.** The official timestep block is replaced by the trained sigma grid (video and audio shifts
179
+ included); the denoising loop hands each step's endpoints to the embedder. Nothing else in the workflow changes.
180
+ - **One file, three workflows.** `t2va` and `fl2va` share `transformer/`. `ref2va` runs the same file on
181
+ `transformer_ref/`, which works because the adapter never touches `adaln_proj`, the carrier of the reference
182
+ conditioning.
183
+
184
+ **File format.** A plain `safetensors` file: PEFT-style keys (`transformer.<module>.lora_A.weight`; bf16 for the
185
+ DiT blocks, fp32 for the two time embedders) plus a self-describing header (`hyperflow_sigmas`, `hyperflow_gate`,
186
+ `lora_rank`, `lora_alpha`, `base_model`, `base_model_revision`, …):
187
+
188
+ ```python
189
+ from hyperflow_h3 import read_metadata
190
+
191
+ meta = read_metadata("minimax_h3_hyperflow_8step_v1.0.safetensors") # header only, no tensors loaded
192
+ meta.base_model_revision, meta.sigmas, meta.gate
193
+ ```
194
+
195
+ ## Intended use and limitations
196
+
197
+ - A drop-in speed-up for MiniMax-H3 inference on diffusers: same inputs, same outputs (video + stereo audio), ~3x
198
+ faster end to end. It is a distilled few-step sampler, so compare against the base Diffusers pipeline on your own
199
+ prompts before adopting it (`--baseline` in the example scripts runs both on the same seed).
200
+ - Prompt following, languages, biases and failure modes are the base model's, which the distillation is designed to
201
+ keep ([Quality](#quality)); see the [MiniMax-H3 model card](https://huggingface.co/MiniMaxAI/MiniMax-H3).
202
+ - Training code is not part of this release.
203
+ - Use is bound by the license below, including its territorial scope and its Acceptable Use Policy.
204
+
205
+ ## License
206
+
207
+ The LoRA weights are a Model Derivative of MiniMax-H3 and are distributed under the
208
+ [MiniMax H3 Community License Agreement](LICENSE), the base model's license, with no additional terms. By
209
+ downloading or using them you accept that Agreement. In particular:
210
+
211
+ - **Territory.** The license is granted worldwide **excluding the European Union, the United Kingdom, the Republic of
212
+ Korea and the United States of America**. Individuals and organizations there must obtain MiniMax's authorization
213
+ before downloading ([application form](https://platform.minimax.io/h3-license), open to individuals as well;
214
+ [Q&A](https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/docs/QA-about-License.md)).
215
+ - **Redistribution.** Pass on a copy of the Agreement and the [`NOTICE`](NOTICE) file, mark modified files, and do not
216
+ impose additional or different terms (Section III).
217
+ - **Disclosure.** Content generated with these weights and posted publicly must be clearly disclosed as AI-generated
218
+ (Acceptable Use Policy, item 12). Naming "MiniMax H3" is encouraged, not required.
219
+ - **Commercial use.** Products or services with more than 20 million US dollars yearly revenue need MiniMax's prior
220
+ written authorization (api@minimax.io). Products and services built with the weights must display "MiniMax H3" on
221
+ their user interface (Section IV.2); this applies to what you build, not to finished videos published on
222
+ third-party platforms.
223
+ - **Hosted services.** If you serve generations to third parties, you must implement the safeguards in Section V.5
224
+ and bind your users to the same use restrictions (Section V.2).
225
+ - **Use restrictions.** Section V and the Acceptable Use Policy (Exhibit A) apply to the weights and to their outputs;
226
+ outputs may not be used to improve other AI models.
227
+ - **Upstream text.** [`LICENSE`](LICENSE) is the August 2, 2026 text. MiniMax may update the Acceptable Use Policy;
228
+ the version in the [upstream repository](https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/LICENSE) governs.
229
+
230
+ The loader and example scripts on [GitHub](https://github.com/Video-Rebirth/hyperflow) are separate software under
231
+ the Apache License 2.0.
232
+
233
+ HyperFlow is developed by Video Rebirth. It is not an official MiniMax product and is not affiliated with or
234
+ endorsed by MiniMax. Powered by MiniMax H3.
235
+
236
+ ## Acknowledgements
237
+
238
+ - **[MiniMax-H3](https://huggingface.co/MiniMaxAI/MiniMax-H3)** ([GitHub](https://github.com/MiniMax-AI/MiniMax-H3)): the base model, VAEs, conditioner and official workflows.
239
+ - **AnyFlow** ([GitHub](https://github.com/NVlabs/AnyFlow); Gu et al., 2026): the flow-map formulation behind the two-time `(t, r)` conditioning.
240
+ - **[diffusers](https://github.com/huggingface/diffusers)**: the Modular Pipeline and context parallel plan the loader builds on.
241
+ - **FlashAttention-3** and **Sol-Attn** ([GitHub](https://github.com/NVlabs/Sana/tree/main/techniques/sparse_backends); Li et al., 2026): optional attention kernels.
242
+
243
+ Thanks to their authors.
hyperflow.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "default": "minimax_h3_hyperflow_8step_v1.0.safetensors",
3
+ "weights": {
4
+ "minimax_h3_hyperflow_8step_v1.0.safetensors": {
5
+ "hyperflow_version": "1.0",
6
+ "sha256": "9297f4505bfdef59c3014d11274411809c19b0abfe26161cab2b425a696df447"
7
+ }
8
+ }
9
+ }
minimax_h3_hyperflow_8step_v1.0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9297f4505bfdef59c3014d11274411809c19b0abfe26161cab2b425a696df447
3
+ size 2795328008