Image-Text-to-Video
MiniMax H3
Diffusion Single File
English
Chinese
ref2va
comfyui
int8-convrot
synchronized-audio-video
experimental
Instructions to use Wesley1234/minimax_h3_ref2va_patchin_hf102 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use Wesley1234/minimax_h3_ref2va_patchin_hf102 with Diffusion Single File:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Upload evidence/metrics_seed_2608103502.json with huggingface_hub
Browse files
evidence/metrics_seed_2608103502.json
ADDED
|
@@ -0,0 +1,343 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"controls": {
|
| 3 |
+
"seed": 2608103502,
|
| 4 |
+
"steps": 20,
|
| 5 |
+
"size": [
|
| 6 |
+
512,
|
| 7 |
+
512
|
| 8 |
+
],
|
| 9 |
+
"frames": 22,
|
| 10 |
+
"prompt": "centered front-facing portrait; identical for all candidates",
|
| 11 |
+
"baseline": "minimax_h3_ref2va_int8_convrot.safetensors"
|
| 12 |
+
},
|
| 13 |
+
"baseline_regions": {
|
| 14 |
+
"forehead": {
|
| 15 |
+
"gradient": 19.118993974634726,
|
| 16 |
+
"laplacian": 16.466526990591827,
|
| 17 |
+
"highpass": 0.8236231262033636
|
| 18 |
+
},
|
| 19 |
+
"left_cheek": {
|
| 20 |
+
"gradient": 20.42510201739831,
|
| 21 |
+
"laplacian": 26.98414471590909,
|
| 22 |
+
"highpass": 1.0676020979881287
|
| 23 |
+
},
|
| 24 |
+
"right_cheek": {
|
| 25 |
+
"gradient": 34.39876539933298,
|
| 26 |
+
"laplacian": 207.0204395568182,
|
| 27 |
+
"highpass": 2.35250105641105
|
| 28 |
+
},
|
| 29 |
+
"portrait": {
|
| 30 |
+
"gradient": 29.842277256068837,
|
| 31 |
+
"laplacian": 86.62322641690345,
|
| 32 |
+
"highpass": 1.685757739977403
|
| 33 |
+
}
|
| 34 |
+
},
|
| 35 |
+
"results": {
|
| 36 |
+
"patchin_hf102": {
|
| 37 |
+
"label": "Patch input 1.02",
|
| 38 |
+
"frame_shape": [
|
| 39 |
+
22,
|
| 40 |
+
512,
|
| 41 |
+
512,
|
| 42 |
+
3
|
| 43 |
+
],
|
| 44 |
+
"rgb_mae_255": 14.241889953613281,
|
| 45 |
+
"psnr_db": 20.817959207130038,
|
| 46 |
+
"parity_abs_diff": [
|
| 47 |
+
14.241340637207031,
|
| 48 |
+
14.256017684936523,
|
| 49 |
+
14.226961135864258,
|
| 50 |
+
14.243230819702148
|
| 51 |
+
],
|
| 52 |
+
"parity_relative_spread": 0.00204021755758276,
|
| 53 |
+
"region_ratios": {
|
| 54 |
+
"forehead": {
|
| 55 |
+
"gradient": 1.032062507677407,
|
| 56 |
+
"laplacian": 1.1970488793243081,
|
| 57 |
+
"highpass": 1.038160105570075
|
| 58 |
+
},
|
| 59 |
+
"left_cheek": {
|
| 60 |
+
"gradient": 1.0283841064243358,
|
| 61 |
+
"laplacian": 0.9345722788494126,
|
| 62 |
+
"highpass": 0.9861274339583423
|
| 63 |
+
},
|
| 64 |
+
"right_cheek": {
|
| 65 |
+
"gradient": 1.036924744809861,
|
| 66 |
+
"laplacian": 1.0964563953998219,
|
| 67 |
+
"highpass": 1.0323794535268038
|
| 68 |
+
},
|
| 69 |
+
"portrait": {
|
| 70 |
+
"gradient": 1.1443190568533455,
|
| 71 |
+
"laplacian": 1.2320932420919597,
|
| 72 |
+
"highpass": 1.193404968594918
|
| 73 |
+
}
|
| 74 |
+
},
|
| 75 |
+
"skin_mean_ratios": {
|
| 76 |
+
"gradient": 1.0324571196372012,
|
| 77 |
+
"laplacian": 1.0760258511911809,
|
| 78 |
+
"highpass": 1.018888997685074
|
| 79 |
+
},
|
| 80 |
+
"skin_highpass_positive_regions": 2,
|
| 81 |
+
"temporal_ratio": 1.0321481634935248,
|
| 82 |
+
"audio_correlation": 0.9930406616019274,
|
| 83 |
+
"audio_rms_ratio": 0.9189266010030337,
|
| 84 |
+
"audio_samples_compared": 26624,
|
| 85 |
+
"audio_rate": 32000
|
| 86 |
+
},
|
| 87 |
+
"patchio_hf101": {
|
| 88 |
+
"label": "Patch I/O 1.01",
|
| 89 |
+
"frame_shape": [
|
| 90 |
+
22,
|
| 91 |
+
512,
|
| 92 |
+
512,
|
| 93 |
+
3
|
| 94 |
+
],
|
| 95 |
+
"rgb_mae_255": 26.415775299072266,
|
| 96 |
+
"psnr_db": 15.82615784912672,
|
| 97 |
+
"parity_abs_diff": [
|
| 98 |
+
26.458765029907227,
|
| 99 |
+
26.356605529785156,
|
| 100 |
+
26.47501564025879,
|
| 101 |
+
26.372785568237305
|
| 102 |
+
],
|
| 103 |
+
"parity_relative_spread": 0.004482549917521291,
|
| 104 |
+
"region_ratios": {
|
| 105 |
+
"forehead": {
|
| 106 |
+
"gradient": 0.9444839764533365,
|
| 107 |
+
"laplacian": 0.8272422765181436,
|
| 108 |
+
"highpass": 0.9609261260659551
|
| 109 |
+
},
|
| 110 |
+
"left_cheek": {
|
| 111 |
+
"gradient": 0.9498798623889819,
|
| 112 |
+
"laplacian": 0.7533049208713888,
|
| 113 |
+
"highpass": 0.9362137479998064
|
| 114 |
+
},
|
| 115 |
+
"right_cheek": {
|
| 116 |
+
"gradient": 0.8615758885295116,
|
| 117 |
+
"laplacian": 0.6397598914467637,
|
| 118 |
+
"highpass": 0.7811897209359578
|
| 119 |
+
},
|
| 120 |
+
"portrait": {
|
| 121 |
+
"gradient": 0.9699411210561822,
|
| 122 |
+
"laplacian": 0.7818911547345777,
|
| 123 |
+
"highpass": 0.9346941424916256
|
| 124 |
+
}
|
| 125 |
+
},
|
| 126 |
+
"skin_mean_ratios": {
|
| 127 |
+
"gradient": 0.91864657579061,
|
| 128 |
+
"laplacian": 0.7401023629454321,
|
| 129 |
+
"highpass": 0.8927765316672397
|
| 130 |
+
},
|
| 131 |
+
"skin_highpass_positive_regions": 0,
|
| 132 |
+
"temporal_ratio": 0.7695742329812562,
|
| 133 |
+
"audio_correlation": 0.6777213227155774,
|
| 134 |
+
"audio_rms_ratio": 0.9347905590619777,
|
| 135 |
+
"audio_samples_compared": 26624,
|
| 136 |
+
"audio_rate": 32000
|
| 137 |
+
},
|
| 138 |
+
"attn_last8_g102_backupAB": {
|
| 139 |
+
"label": "Attention 1.02",
|
| 140 |
+
"frame_shape": [
|
| 141 |
+
22,
|
| 142 |
+
512,
|
| 143 |
+
512,
|
| 144 |
+
3
|
| 145 |
+
],
|
| 146 |
+
"rgb_mae_255": 24.50489044189453,
|
| 147 |
+
"psnr_db": 16.033918404344604,
|
| 148 |
+
"parity_abs_diff": [
|
| 149 |
+
24.49074935913086,
|
| 150 |
+
24.49752426147461,
|
| 151 |
+
24.511693954467773,
|
| 152 |
+
24.51958656311035
|
| 153 |
+
],
|
| 154 |
+
"parity_relative_spread": 0.0011767939257850033,
|
| 155 |
+
"region_ratios": {
|
| 156 |
+
"forehead": {
|
| 157 |
+
"gradient": 0.9651551361715502,
|
| 158 |
+
"laplacian": 0.8632519050145561,
|
| 159 |
+
"highpass": 0.9715350187455672
|
| 160 |
+
},
|
| 161 |
+
"left_cheek": {
|
| 162 |
+
"gradient": 0.895689344936705,
|
| 163 |
+
"laplacian": 0.6889516854452177,
|
| 164 |
+
"highpass": 0.8737009644258836
|
| 165 |
+
},
|
| 166 |
+
"right_cheek": {
|
| 167 |
+
"gradient": 0.9698874011966927,
|
| 168 |
+
"laplacian": 0.8081264950412943,
|
| 169 |
+
"highpass": 0.8652700024587207
|
| 170 |
+
},
|
| 171 |
+
"portrait": {
|
| 172 |
+
"gradient": 0.9997757169194811,
|
| 173 |
+
"laplacian": 1.0655103462302664,
|
| 174 |
+
"highpass": 0.9821616419372078
|
| 175 |
+
}
|
| 176 |
+
},
|
| 177 |
+
"skin_mean_ratios": {
|
| 178 |
+
"gradient": 0.9435772941016493,
|
| 179 |
+
"laplacian": 0.7867766951670226,
|
| 180 |
+
"highpass": 0.9035019952100573
|
| 181 |
+
},
|
| 182 |
+
"skin_highpass_positive_regions": 0,
|
| 183 |
+
"temporal_ratio": 0.9440205176611185,
|
| 184 |
+
"audio_correlation": 0.6593426872963273,
|
| 185 |
+
"audio_rms_ratio": 0.976732972125702,
|
| 186 |
+
"audio_samples_compared": 26624,
|
| 187 |
+
"audio_rate": 32000
|
| 188 |
+
},
|
| 189 |
+
"attn_last8_g105_backupAB": {
|
| 190 |
+
"label": "Attention 1.05",
|
| 191 |
+
"frame_shape": [
|
| 192 |
+
22,
|
| 193 |
+
512,
|
| 194 |
+
512,
|
| 195 |
+
3
|
| 196 |
+
],
|
| 197 |
+
"rgb_mae_255": 26.17381477355957,
|
| 198 |
+
"psnr_db": 15.466375766270218,
|
| 199 |
+
"parity_abs_diff": [
|
| 200 |
+
26.163225173950195,
|
| 201 |
+
26.164703369140625,
|
| 202 |
+
26.182741165161133,
|
| 203 |
+
26.184694290161133
|
| 204 |
+
],
|
| 205 |
+
"parity_relative_spread": 0.000820250883745451,
|
| 206 |
+
"region_ratios": {
|
| 207 |
+
"forehead": {
|
| 208 |
+
"gradient": 0.9959880421304674,
|
| 209 |
+
"laplacian": 0.8540994665054672,
|
| 210 |
+
"highpass": 0.988992733340178
|
| 211 |
+
},
|
| 212 |
+
"left_cheek": {
|
| 213 |
+
"gradient": 0.9608255294610666,
|
| 214 |
+
"laplacian": 0.8020721975955389,
|
| 215 |
+
"highpass": 0.9816722853274565
|
| 216 |
+
},
|
| 217 |
+
"right_cheek": {
|
| 218 |
+
"gradient": 0.9779253911003554,
|
| 219 |
+
"laplacian": 0.7476207497429617,
|
| 220 |
+
"highpass": 0.8529531788916929
|
| 221 |
+
},
|
| 222 |
+
"portrait": {
|
| 223 |
+
"gradient": 1.004480955684384,
|
| 224 |
+
"laplacian": 0.9810051504775767,
|
| 225 |
+
"highpass": 0.9774153908587273
|
| 226 |
+
}
|
| 227 |
+
},
|
| 228 |
+
"skin_mean_ratios": {
|
| 229 |
+
"gradient": 0.9782463208972964,
|
| 230 |
+
"laplacian": 0.8012641379479893,
|
| 231 |
+
"highpass": 0.9412060658531091
|
| 232 |
+
},
|
| 233 |
+
"skin_highpass_positive_regions": 0,
|
| 234 |
+
"temporal_ratio": 0.7998909601840626,
|
| 235 |
+
"audio_correlation": 0.7073727262310824,
|
| 236 |
+
"audio_rms_ratio": 1.0813777891956717,
|
| 237 |
+
"audio_samples_compared": 26624,
|
| 238 |
+
"audio_rate": 32000
|
| 239 |
+
},
|
| 240 |
+
"adaln_video_g099": {
|
| 241 |
+
"label": "Video AdaLN 0.99",
|
| 242 |
+
"frame_shape": [
|
| 243 |
+
22,
|
| 244 |
+
512,
|
| 245 |
+
512,
|
| 246 |
+
3
|
| 247 |
+
],
|
| 248 |
+
"rgb_mae_255": 9.915425300598145,
|
| 249 |
+
"psnr_db": 23.4676943237718,
|
| 250 |
+
"parity_abs_diff": [
|
| 251 |
+
9.92910385131836,
|
| 252 |
+
9.926578521728516,
|
| 253 |
+
9.905007362365723,
|
| 254 |
+
9.90101146697998
|
| 255 |
+
],
|
| 256 |
+
"parity_relative_spread": 0.0028332001388467164,
|
| 257 |
+
"region_ratios": {
|
| 258 |
+
"forehead": {
|
| 259 |
+
"gradient": 0.9991286249222505,
|
| 260 |
+
"laplacian": 1.0026000155229857,
|
| 261 |
+
"highpass": 1.006577991424609
|
| 262 |
+
},
|
| 263 |
+
"left_cheek": {
|
| 264 |
+
"gradient": 0.9780371131275568,
|
| 265 |
+
"laplacian": 0.9120942549436786,
|
| 266 |
+
"highpass": 0.9648077761342945
|
| 267 |
+
},
|
| 268 |
+
"right_cheek": {
|
| 269 |
+
"gradient": 1.0072948081898199,
|
| 270 |
+
"laplacian": 0.978957933377793,
|
| 271 |
+
"highpass": 1.0015376344560063
|
| 272 |
+
},
|
| 273 |
+
"portrait": {
|
| 274 |
+
"gradient": 0.9180274832210044,
|
| 275 |
+
"laplacian": 0.8724958236770269,
|
| 276 |
+
"highpass": 0.8997279055195478
|
| 277 |
+
}
|
| 278 |
+
},
|
| 279 |
+
"skin_mean_ratios": {
|
| 280 |
+
"gradient": 0.9948201820798758,
|
| 281 |
+
"laplacian": 0.9645507346148191,
|
| 282 |
+
"highpass": 0.9909744673383032
|
| 283 |
+
},
|
| 284 |
+
"skin_highpass_positive_regions": 2,
|
| 285 |
+
"temporal_ratio": 0.9975294373750192,
|
| 286 |
+
"audio_correlation": 0.8806014978763488,
|
| 287 |
+
"audio_rms_ratio": 0.9717682054851432,
|
| 288 |
+
"audio_samples_compared": 26624,
|
| 289 |
+
"audio_rate": 32000
|
| 290 |
+
},
|
| 291 |
+
"adaln_video_g101": {
|
| 292 |
+
"label": "Video AdaLN 1.01",
|
| 293 |
+
"frame_shape": [
|
| 294 |
+
22,
|
| 295 |
+
512,
|
| 296 |
+
512,
|
| 297 |
+
3
|
| 298 |
+
],
|
| 299 |
+
"rgb_mae_255": 21.380828857421875,
|
| 300 |
+
"psnr_db": 16.910562972665033,
|
| 301 |
+
"parity_abs_diff": [
|
| 302 |
+
21.360219955444336,
|
| 303 |
+
21.358842849731445,
|
| 304 |
+
21.403270721435547,
|
| 305 |
+
21.400985717773438
|
| 306 |
+
],
|
| 307 |
+
"parity_relative_spread": 0.002077930187772434,
|
| 308 |
+
"region_ratios": {
|
| 309 |
+
"forehead": {
|
| 310 |
+
"gradient": 1.0077906715102303,
|
| 311 |
+
"laplacian": 0.9818674456823687,
|
| 312 |
+
"highpass": 1.0348544321311723
|
| 313 |
+
},
|
| 314 |
+
"left_cheek": {
|
| 315 |
+
"gradient": 0.9018172697800978,
|
| 316 |
+
"laplacian": 0.6945614191520033,
|
| 317 |
+
"highpass": 0.8807424179496622
|
| 318 |
+
},
|
| 319 |
+
"right_cheek": {
|
| 320 |
+
"gradient": 0.9685711252044322,
|
| 321 |
+
"laplacian": 0.8277841438265235,
|
| 322 |
+
"highpass": 0.8748947150154358
|
| 323 |
+
},
|
| 324 |
+
"portrait": {
|
| 325 |
+
"gradient": 1.0097364352355103,
|
| 326 |
+
"laplacian": 1.0435720605395844,
|
| 327 |
+
"highpass": 0.9943788850381092
|
| 328 |
+
}
|
| 329 |
+
},
|
| 330 |
+
"skin_mean_ratios": {
|
| 331 |
+
"gradient": 0.9593930221649201,
|
| 332 |
+
"laplacian": 0.8347376695536318,
|
| 333 |
+
"highpass": 0.9301638550320902
|
| 334 |
+
},
|
| 335 |
+
"skin_highpass_positive_regions": 1,
|
| 336 |
+
"temporal_ratio": 1.0230604475597387,
|
| 337 |
+
"audio_correlation": 0.8226417228875448,
|
| 338 |
+
"audio_rms_ratio": 0.9860379990897791,
|
| 339 |
+
"audio_samples_compared": 26624,
|
| 340 |
+
"audio_rate": 32000
|
| 341 |
+
}
|
| 342 |
+
}
|
| 343 |
+
}
|