Model save
Browse files
README.md
ADDED
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
base_model: microsoft/swin-tiny-patch4-window7-224
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- imagefolder
|
8 |
+
metrics:
|
9 |
+
- accuracy
|
10 |
+
- precision
|
11 |
+
- recall
|
12 |
+
- f1
|
13 |
+
model-index:
|
14 |
+
- name: batch-size16_Celeb-DF-v2_opencv-1FPS_faces-expand30-aligned_unaugmentation
|
15 |
+
results:
|
16 |
+
- task:
|
17 |
+
name: Image Classification
|
18 |
+
type: image-classification
|
19 |
+
dataset:
|
20 |
+
name: imagefolder
|
21 |
+
type: imagefolder
|
22 |
+
config: default
|
23 |
+
split: test
|
24 |
+
args: default
|
25 |
+
metrics:
|
26 |
+
- name: Accuracy
|
27 |
+
type: accuracy
|
28 |
+
value: 0.9931273636466434
|
29 |
+
- name: Precision
|
30 |
+
type: precision
|
31 |
+
value: 0.994805662520431
|
32 |
+
- name: Recall
|
33 |
+
type: recall
|
34 |
+
value: 0.9976108124626689
|
35 |
+
- name: F1
|
36 |
+
type: f1
|
37 |
+
value: 0.9962062627873912
|
38 |
+
---
|
39 |
+
|
40 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
41 |
+
should probably proofread and complete it, then remove this comment. -->
|
42 |
+
|
43 |
+
# batch-size16_Celeb-DF-v2_opencv-1FPS_faces-expand30-aligned_unaugmentation
|
44 |
+
|
45 |
+
This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
|
46 |
+
It achieves the following results on the evaluation set:
|
47 |
+
- Loss: 0.0198
|
48 |
+
- Accuracy: 0.9931
|
49 |
+
- Precision: 0.9948
|
50 |
+
- Recall: 0.9976
|
51 |
+
- F1: 0.9962
|
52 |
+
- Roc Auc: 0.9991
|
53 |
+
|
54 |
+
## Model description
|
55 |
+
|
56 |
+
More information needed
|
57 |
+
|
58 |
+
## Intended uses & limitations
|
59 |
+
|
60 |
+
More information needed
|
61 |
+
|
62 |
+
## Training and evaluation data
|
63 |
+
|
64 |
+
More information needed
|
65 |
+
|
66 |
+
## Training procedure
|
67 |
+
|
68 |
+
### Training hyperparameters
|
69 |
+
|
70 |
+
The following hyperparameters were used during training:
|
71 |
+
- learning_rate: 5e-05
|
72 |
+
- train_batch_size: 16
|
73 |
+
- eval_batch_size: 16
|
74 |
+
- seed: 42
|
75 |
+
- gradient_accumulation_steps: 4
|
76 |
+
- total_train_batch_size: 64
|
77 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
78 |
+
- lr_scheduler_type: linear
|
79 |
+
- lr_scheduler_warmup_ratio: 0.1
|
80 |
+
- num_epochs: 1
|
81 |
+
|
82 |
+
### Training results
|
83 |
+
|
84 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Roc Auc |
|
85 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:-------:|
|
86 |
+
| 0.0338 | 0.9994 | 1243 | 0.0198 | 0.9931 | 0.9948 | 0.9976 | 0.9962 | 0.9991 |
|
87 |
+
|
88 |
+
|
89 |
+
### Framework versions
|
90 |
+
|
91 |
+
- Transformers 4.41.2
|
92 |
+
- Pytorch 2.3.1
|
93 |
+
- Datasets 2.20.0
|
94 |
+
- Tokenizers 0.19.1
|