File size: 2,395 Bytes
3b6f2f0 62b4a02 3b6f2f0 c98ba92 3b6f2f0 c98ba92 3b6f2f0 c98ba92 3b6f2f0 c98ba92 3b6f2f0 c98ba92 3b6f2f0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
---
library_name: keras
widget:
- text: input
output:
url: ./assets/input.png
- text: target
output:
url: ./assets/target.png
- text: output
output:
url: ./assets/output.png
tags:
- adversarial
- aesthetic
- quality
- filter
metrics:
- TopIQ-FR
- ArcFace Cosine Distance
license: gpl-3.0
base_model:
- vnet
- logasja/ArcFace
pipeline_tag: image-to-image
datasets:
- logasja/FDF
---
<Gallery />
Training logs [here](https://wandb.ai/spuds/auramask/runs/c7f3f8c1813cc1ea9727ea45501e3ce1)
# Model Description
This model uses a modified vnet for 2D input/output implemented [here](https://github.com/logasja/keras3-unets) with the following configuration.
```json
{
"activation": "ReLU",
"batch_norm": false,
"filter_num": [
64,
128,
256,
512,
512
],
"n_labels": 3,
"output_activation": "tanh",
"pool": false,
"res_num_ini": 1,
"res_num_max": 3,
"unpool": false
}
```
```json
{
"alpha": 0.0001,
"batch": 64,
"epochs": 500,
"epsilon": 1,
"input": "(256, 256)",
"losses": {
"FEAT_ArcFace": {
"d": "cosine_similarity",
"f": "ArcFace",
"name": "FEAT_ArcFace",
"reduction": "sum_over_batch_size",
"threshold": 0.68,
"weight": 0.1
},
"TopIQ": {
"full_ref": true,
"lower_better": false,
"name": "TopIQ",
"reduction": "sum_over_batch_size",
"score_range": "~0, ~1",
"weight": 0.9
},
"mean_squared_error": {
"name": "mean_squared_error",
"reduction": "sum_over_batch_size",
"weight": 0.1
}
},
"mixed_precision": true,
"optimizer": {
"amsgrad": false,
"beta_1": 0.9,
"beta_2": 0.999,
"clipnorm": null,
"clipvalue": null,
"ema_momentum": 0.99,
"ema_overwrite_frequency": null,
"epsilon": 1e-07,
"global_clipnorm": null,
"gradient_accumulation_steps": null,
"learning_rate": 9.999999747378752e-05,
"loss_scale_factor": null,
"name": "adamw",
"use_ema": false,
"weight_decay": 0.004
},
"seed": "BIIIIIGSTRETCH",
"testing": 0.01,
"training": 0.99
}
```
## Model Architecture Plot
 |