Hugging Face's logo

MiniMaxAI
/
MiniMax-H3

Image-Text-to-Video
Diffusers
Safetensors
MiniMaxH3ModularPipeline
text-to-video
image-to-video
video-to-video
text-to-audio-video
image-to-audio-video
image-text-to-audio-video
video-to-audio-video
audio-to-audio-video
audio-video-generation
multimodal
synchronized-audio-video
reference-to-audio-video
Model card Files
xet
 Community
38

Scope the download command to the original checkpoint folders

#16
by multimodalart HF Staff - opened
base: refs/heads/main
←
from: refs/pr/16
Discussion Files changed
+69
-143
Files changed (20) hide show
  1. README.md +2 -4
  2. assets/action-reference.mov +3 -0
  3. assets/character-action-reference.png +3 -0
  4. assets/character-replacement-action-reference.mp4 +3 -0
  5. assets/fashion-glasses-ad.mp4 +3 -0
  6. assets/fashion-glasses-reference-1.png +3 -0
  7. assets/fashion-glasses-reference-2.png +3 -0
  8. assets/fashion-glasses-reference-3.png +3 -0
  9. assets/fashion-glasses-reference-4.png +3 -0
  10. assets/fl2va-clay-fox-reference.png +3 -0
  11. assets/fl2va-clay-fox.mp4 +3 -0
  12. assets/h3-architecture.png +3 -0
  13. assets/h3-cinematic-shot.mp4 +3 -0
  14. assets/h3-suspense-title.mp4 +3 -0
  15. assets/logo.svg +16 -0
  16. assets/reference-image-1.png +3 -0
  17. assets/reference-image-2.png +3 -0
  18. assets/robot-arm-red-cube.mp4 +3 -0
  19. assets/t2va-768p-demo.mp4 +3 -0
  20. model_index.json +0 -139
README.md CHANGED
@@ -194,14 +194,12 @@ Each checkpoint is distributed as a self\-contained Hugging Face\-style reposito
194
 
195
  Download the model. The repository hosts the original checkpoint (`FL2VA/`, `Ref2VA/`) and the diffusers format side by side, so scope the download to what your framework needs:
196
 
197
- `model_index.json` is the repository-level modular index. The task-family-specific diffusers indexes remain under `FL2VA/model_index.json` and `Ref2VA/model_index.json`.
198
-
199
  ```bash
200
  # Original checkpoint, both task families (SGLang, vLLM):
201
- hf download MiniMaxAI/MiniMax-H3 --include "model_index.json" "modular_model_index.json" "FL2VA/*" "Ref2VA/*" --local-dir MiniMax-H3
202
 
203
  # Or a single task family:
204
- hf download MiniMaxAI/MiniMax-H3 --include "model_index.json" "modular_model_index.json" "FL2VA/*" --local-dir MiniMax-H3
205
  ```
206
 
207
  diffusers users do not need a manual download: `ModularPipeline.from_pretrained("MiniMaxAI/MiniMax-H3")` fetches exactly the components it needs. See the [diffusers documentation](https://github.com/huggingface/diffusers/blob/minimax-h3/docs/source/en/api/pipelines/minimax_h3.md) for loading recipes.
 
194
 
195
  Download the model. The repository hosts the original checkpoint (`FL2VA/`, `Ref2VA/`) and the diffusers format side by side, so scope the download to what your framework needs:
196
 
 
 
197
  ```bash
198
  # Original checkpoint, both task families (SGLang, vLLM):
199
+ hf download MiniMaxAI/MiniMax-H3 --include "FL2VA/*" "Ref2VA/*" --local-dir MiniMax-H3
200
 
201
  # Or a single task family:
202
+ hf download MiniMaxAI/MiniMax-H3 --include "FL2VA/*" --local-dir MiniMax-H3
203
  ```
204
 
205
  diffusers users do not need a manual download: `ModularPipeline.from_pretrained("MiniMaxAI/MiniMax-H3")` fetches exactly the components it needs. See the [diffusers documentation](https://github.com/huggingface/diffusers/blob/minimax-h3/docs/source/en/api/pipelines/minimax_h3.md) for loading recipes.
assets/action-reference.mov ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:116e0f08a399834e7ffc3472d036659b33250f4ba4f0b7e63f17ec07cb58e4dc
3
+ size 31446151
assets/character-action-reference.png ADDED

Git LFS Details

  • SHA256: 01a6c6f29e1ce9d414276e6e2eca06b171e5a68dd54a86074a4ab77bb9cb8883
  • Pointer size: 132 Bytes
  • Size of remote file: 5.24 MB
assets/character-replacement-action-reference.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:796c264126110d87992dcb54213ac0697920cb4ddf3d5a06aa36069386fc4fa1
3
+ size 8223683
assets/fashion-glasses-ad.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a11e6b0095dd7766b724bb886edf4d7d9930af44da6839ee195269ac4fc60ba4
3
+ size 10938085
assets/fashion-glasses-reference-1.png ADDED

Git LFS Details

  • SHA256: 5665f6529dd73339c0ac9397f0bd52df0cd44bd5b63d68af88f95855879d3d8b
  • Pointer size: 132 Bytes
  • Size of remote file: 1.01 MB
assets/fashion-glasses-reference-2.png ADDED

Git LFS Details

  • SHA256: 3a74bdab0bff0e126aae3757833fd2ec8cfba9afcb2ca36f63ee86977f7f7c00
  • Pointer size: 131 Bytes
  • Size of remote file: 889 kB
assets/fashion-glasses-reference-3.png ADDED

Git LFS Details

  • SHA256: 7b1491e5a227f226fa1395e71ac89a84358a435d974acb9f9065b9ded0982cbe
  • Pointer size: 132 Bytes
  • Size of remote file: 1.06 MB
assets/fashion-glasses-reference-4.png ADDED

Git LFS Details

  • SHA256: c9db6d6788855029ef9ac4109b05e17693ddb3bdf6ae3f24cec44de478c00f13
  • Pointer size: 131 Bytes
  • Size of remote file: 368 kB
assets/fl2va-clay-fox-reference.png ADDED

Git LFS Details

  • SHA256: 55eff74469b79a100e63bdd358596f6cff671e18fca9b276d6faef16faf22d27
  • Pointer size: 132 Bytes
  • Size of remote file: 1.73 MB
assets/fl2va-clay-fox.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4604b0838b736ecaa209c978f880bfacbb3a0fc82a8a0517c1f5aa16454b50ef
3
+ size 10931936
assets/h3-architecture.png ADDED

Git LFS Details

  • SHA256: 9765612d331f5fd31068a9283ffe28f11963f6a418c9f4093f81b408ea756630
  • Pointer size: 131 Bytes
  • Size of remote file: 336 kB
assets/h3-cinematic-shot.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:436defc81cfa7d53aef423f368be82ead20056088087e0be308b6e3865e8fb81
3
+ size 4361547
assets/h3-suspense-title.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:90ebbd7edc71c9a0151c3064126acd2dbb5b7cba58ce459e223e29f2c04f9186
3
+ size 16809220
assets/logo.svg ADDED
assets/reference-image-1.png ADDED

Git LFS Details

  • SHA256: 49b89115da4e253228969830ca491e4ca1b79331eb241a905d1ebca3dc6de7be
  • Pointer size: 131 Bytes
  • Size of remote file: 721 kB
assets/reference-image-2.png ADDED

Git LFS Details

  • SHA256: c9db6d6788855029ef9ac4109b05e17693ddb3bdf6ae3f24cec44de478c00f13
  • Pointer size: 131 Bytes
  • Size of remote file: 368 kB
assets/robot-arm-red-cube.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a751e6100dbf6502a99a2adc0b12171304ca7b17eecea70e2ed035a73ea692c
3
+ size 2259732
assets/t2va-768p-demo.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d66903241362e224085cc93f7a5e70fba6ab378d0ac6ff6af83acf2559849a42
3
+ size 1637373
model_index.json DELETED
@@ -1,139 +0,0 @@
1
- {
2
- "_class_name": "MiniMaxH3ModularPipeline",
3
- "_diffusers_version": "0.36.0.dev0",
4
- "_blocks_class_name": "MiniMaxH3Blocks",
5
- "_minimax_h3": {
6
- "schema_version": 1,
7
- "index_scope": "repository",
8
- "task_family_indexes": {
9
- "fl2va": "FL2VA/model_index.json",
10
- "ref2va": "Ref2VA/model_index.json"
11
- }
12
- },
13
- "text_encoder": [
14
- "transformers",
15
- "Qwen3VLForConditionalGeneration",
16
- {
17
- "type_hint": [
18
- "transformers",
19
- "Qwen3VLForConditionalGeneration"
20
- ],
21
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
22
- "subfolder": "text_encoder",
23
- "variant": null,
24
- "revision": null
25
- }
26
- ],
27
- "tokenizer": [
28
- "transformers",
29
- "Qwen2TokenizerFast",
30
- {
31
- "type_hint": [
32
- "transformers",
33
- "Qwen2TokenizerFast"
34
- ],
35
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
36
- "subfolder": "tokenizer",
37
- "variant": null,
38
- "revision": null
39
- }
40
- ],
41
- "processor": [
42
- "transformers",
43
- "Qwen3VLProcessor",
44
- {
45
- "type_hint": [
46
- "transformers",
47
- "Qwen3VLProcessor"
48
- ],
49
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
50
- "subfolder": "processor",
51
- "variant": null,
52
- "revision": null
53
- }
54
- ],
55
- "vae": [
56
- "diffusers",
57
- "AutoencoderKLMiniMaxH3",
58
- {
59
- "type_hint": [
60
- "diffusers",
61
- "AutoencoderKLMiniMaxH3"
62
- ],
63
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
64
- "subfolder": "vae",
65
- "variant": null,
66
- "revision": null
67
- }
68
- ],
69
- "audio_vae": [
70
- "diffusers",
71
- "AutoencoderKLMiniMaxH3Audio",
72
- {
73
- "type_hint": [
74
- "diffusers",
75
- "AutoencoderKLMiniMaxH3Audio"
76
- ],
77
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
78
- "subfolder": "audio_vae",
79
- "variant": null,
80
- "revision": null
81
- }
82
- ],
83
- "transformer": [
84
- "diffusers",
85
- "MiniMaxH3Transformer3DModel",
86
- {
87
- "type_hint": [
88
- "diffusers",
89
- "MiniMaxH3Transformer3DModel"
90
- ],
91
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
92
- "subfolder": "transformer",
93
- "variant": null,
94
- "revision": null
95
- }
96
- ],
97
- "transformer_ref": [
98
- "diffusers",
99
- "MiniMaxH3Transformer3DModel",
100
- {
101
- "type_hint": [
102
- "diffusers",
103
- "MiniMaxH3Transformer3DModel"
104
- ],
105
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
106
- "subfolder": "transformer_ref",
107
- "variant": null,
108
- "revision": null
109
- }
110
- ],
111
- "scheduler": [
112
- "diffusers",
113
- "MiniMaxH3Scheduler",
114
- {
115
- "type_hint": [
116
- "diffusers",
117
- "MiniMaxH3Scheduler"
118
- ],
119
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
120
- "subfolder": "scheduler",
121
- "variant": null,
122
- "revision": null
123
- }
124
- ],
125
- "audio_scheduler": [
126
- "diffusers",
127
- "MiniMaxH3Scheduler",
128
- {
129
- "type_hint": [
130
- "diffusers",
131
- "MiniMaxH3Scheduler"
132
- ],
133
- "pretrained_model_name_or_path": "MiniMaxAI/MiniMax-H3",
134
- "subfolder": "audio_scheduler",
135
- "variant": null,
136
- "revision": null
137
- }
138
- ]
139
- }