README.md

---
tags:
- image-classification
- timm
- transformers
- animetimm
- dghs-imgutils
library_name: timm
license: gpl-3.0
datasets:
- animetimm/danbooru-wdtagger-v4-w640-ws-full
base_model:
- timm/resnet152.a1h_in1k
---

# Anime Tagger resnet152.dbv4-full

## Model Details

- **Model Type:** Multilabel Image classification / feature backbone
- **Model Stats:**
  - Params: 83.7M
  - FLOPs / MACs: 67.9G / 33.9G
  - Image size: train = 384 x 384, test = 384 x 384
- **Dataset:** [animetimm/danbooru-wdtagger-v4-w640-ws-full](https://huggingface.co/datasets/animetimm/danbooru-wdtagger-v4-w640-ws-full)
  - Tags Count: 12476
    - General (#0) Tags Count: 9225
    - Character (#4) Tags Count: 3247
    - Rating (#9) Tags Count: 4

## Results

|     #      |    Macro@0.40 (F1/MCC/P/R)    |    Micro@0.40 (F1/MCC/P/R)    |  Macro@Best (F1/P/R)  |
|:----------:|:-----------------------------:|:-----------------------------:|:---------------------:|
| Validation | 0.446 / 0.455 / 0.526 / 0.412 | 0.624 / 0.623 / 0.657 / 0.593 |          ---          |
|    Test    | 0.448 / 0.456 / 0.526 / 0.413 | 0.624 / 0.624 / 0.657 / 0.594 | 0.486 / 0.513 / 0.488 |

* `Macro/Micro@0.40` means the metrics on the threshold 0.40.
* `Macro@Best` means the mean metrics on the tag-level thresholds on each tags, which should have the best F1 scores.

## Thresholds

|  Category  |   Name    |  Alpha  |  Threshold  |  Micro@Thr (F1/P/R)   |  Macro@0.40 (F1/P/R)  |  Macro@Best (F1/P/R)  |
|:----------:|:---------:|:-------:|:-----------:|:---------------------:|:---------------------:|:---------------------:|
|     0      |  general  |    1    |    0.35     | 0.612 / 0.617 / 0.608 | 0.319 / 0.410 / 0.283 | 0.364 / 0.379 / 0.381 |
|     4      | character |    1    |    0.48     | 0.846 / 0.903 / 0.795 | 0.813 / 0.855 / 0.781 | 0.835 / 0.895 / 0.791 |
|     9      |  rating   |    1    |    0.38     | 0.796 / 0.744 / 0.856 | 0.803 / 0.778 / 0.836 | 0.806 / 0.778 / 0.839 |

* `Micro@Thr` means the metrics on the category-level suggested thresholds, which are listed in the table above.
* `Macro@0.40` means the metrics on the threshold 0.40.
* `Macro@Best` means the metrics on the tag-level thresholds on each tags, which should have the best F1 scores.

For tag-level thresholds, you can find them in [selected_tags.csv](https://huggingface.co/animetimm/resnet152.dbv4-full/resolve/main/selected_tags.csv).

## How to Use

We provided a sample image for our code samples, you can find it [here](https://huggingface.co/animetimm/resnet152.dbv4-full/blob/main/sample.webp).

### Use TIMM And Torch

Install [dghs-imgutils](https://github.com/deepghs/imgutils), [timm](https://github.com/huggingface/pytorch-image-models) and other necessary requirements with the following command

```shell
pip install 'dghs-imgutils>=0.17.0' torch huggingface_hub timm pillow pandas
```

After that you can load this model with timm library, and use it for train, validation and test, with the following code

```python
import json

import pandas as pd
import torch
from huggingface_hub import hf_hub_download
from imgutils.data import load_image
from imgutils.preprocess import create_torchvision_transforms
from timm import create_model

repo_id = 'animetimm/resnet152.dbv4-full'
model = create_model(f'hf-hub:{repo_id}', pretrained=True)
model.eval()

with open(hf_hub_download(repo_id=repo_id, repo_type='model', filename='preprocess.json'), 'r') as f:
    preprocessor = create_torchvision_transforms(json.load(f)['test'])
# Compose(
#     PadToSize(size=(512, 512), interpolation=bilinear, background_color=white)
#     Resize(size=384, interpolation=bicubic, max_size=None, antialias=True)
#     CenterCrop(size=[384, 384])
#     MaybeToTensor()
#     Normalize(mean=tensor([0.4850, 0.4560, 0.4060]), std=tensor([0.2290, 0.2240, 0.2250]))
# )

image = load_image('https://huggingface.co/animetimm/resnet152.dbv4-full/resolve/main/sample.webp')
input_ = preprocessor(image).unsqueeze(0)
# input_, shape: torch.Size([1, 3, 384, 384]), dtype: torch.float32
with torch.no_grad():
    output = model(input_)
    prediction = torch.sigmoid(output)[0]
# output, shape: torch.Size([1, 12476]), dtype: torch.float32
# prediction, shape: torch.Size([12476]), dtype: torch.float32

df_tags = pd.read_csv(
    hf_hub_download(repo_id=repo_id, repo_type='model', filename='selected_tags.csv'),
    keep_default_na=False
)
tags = df_tags['name']
mask = prediction.numpy() >= df_tags['best_threshold']
print(dict(zip(tags[mask].tolist(), prediction[mask].tolist())))
# {'general': 0.5816617608070374,
#  'sensitive': 0.4577067196369171,
#  '1girl': 0.9958261251449585,
#  'solo': 0.9633037447929382,
#  'looking_at_viewer': 0.8500308990478516,
#  'blush': 0.8500121831893921,
#  'smile': 0.9115327596664429,
#  'short_hair': 0.7201621532440186,
#  'shirt': 0.6373851299285889,
#  'long_sleeves': 0.8233276605606079,
#  'holding': 0.718224823474884,
#  'dress': 0.5167701840400696,
#  'closed_mouth': 0.5078815221786499,
#  'purple_eyes': 0.6588500142097473,
#  'upper_body': 0.30264830589294434,
#  'flower': 0.9456981420516968,
#  'braid': 0.9781820774078369,
#  'outdoors': 0.3688752353191376,
#  'red_hair': 0.7415601015090942,
#  'blunt_bangs': 0.4880999028682709,
#  'apron': 0.6254206299781799,
#  'plant': 0.33948495984077454,
#  'blue_flower': 0.9264647364616394,
#  'backlighting': 0.14451347291469574,
#  'crown_braid': 0.8123992681503296,
#  'potted_plant': 0.2292894870042801,
#  'flower_pot': 0.29513847827911377,
#  'wiping_tears': 0.46064630150794983}
```
### Use ONNX Model For Inference

Install [dghs-imgutils](https://github.com/deepghs/imgutils) with the following command

```shell
pip install 'dghs-imgutils>=0.17.0'
```

Use `multilabel_timm_predict` function with the following code

```python
from imgutils.generic import multilabel_timm_predict

general, character, rating = multilabel_timm_predict(
    'https://huggingface.co/animetimm/resnet152.dbv4-full/resolve/main/sample.webp',
    repo_id='animetimm/resnet152.dbv4-full',
    fmt=('general', 'character', 'rating'),
)

print(general)
# {'1girl': 0.9958261251449585,
#  'braid': 0.9781820774078369,
#  'solo': 0.9633036851882935,
#  'flower': 0.9456979632377625,
#  'blue_flower': 0.926464855670929,
#  'smile': 0.9115328788757324,
#  'looking_at_viewer': 0.8500310778617859,
#  'blush': 0.8500126600265503,
#  'long_sleeves': 0.8233274221420288,
#  'crown_braid': 0.8123989105224609,
#  'red_hair': 0.7415597438812256,
#  'short_hair': 0.7201613783836365,
#  'holding': 0.7182247638702393,
#  'purple_eyes': 0.6588503122329712,
#  'shirt': 0.6373854279518127,
#  'apron': 0.6254194974899292,
#  'dress': 0.516771137714386,
#  'closed_mouth': 0.5078825354576111,
#  'blunt_bangs': 0.4881010055541992,
#  'wiping_tears': 0.46064335107803345,
#  'outdoors': 0.36887407302856445,
#  'plant': 0.3394850194454193,
#  'upper_body': 0.3026488721370697,
#  'flower_pot': 0.29513871669769287,
#  'potted_plant': 0.22929036617279053,
#  'backlighting': 0.14451327919960022}
print(character)
# {}
print(rating)
# {'general': 0.5816621780395508, 'sensitive': 0.45770588517189026}
```

For further information, see [documentation of function multilabel_timm_predict](https://dghs-imgutils.deepghs.org/main/api_doc/generic/multilabel_timm.html#multilabel-timm-predict).