Update README.md
Browse files
README.md
CHANGED
|
@@ -11,7 +11,7 @@ pipeline_tag: text-to-image
|
|
| 11 |
|
| 12 |
This repository contains model weights and data resources used in the **Type-R** project. The dataset is designed to support text-to-image generation, OCR, text erasing, editing, and evaluation pipelines used in the Type-R system.
|
| 13 |
|
| 14 |
-
## Directory structure
|
| 15 |
|
| 16 |
⚠️ The code in the [repository](https://github.com/CyberAgentAILab/Type-R) is designed to operate directly using this structure.
|
| 17 |
|
|
@@ -45,7 +45,7 @@ resources/
|
|
| 45 |
|
| 46 |
</pre>
|
| 47 |
|
| 48 |
-
## ⚠️
|
| 49 |
### `resources/weight/ocr/solo.pth`
|
| 50 |
Please download this weight from the official [Deeosolo](https://github.com/ViTAE-Transformer/DeepSolo) implementation.
|
| 51 |
[[link](https://onedrive.live.com/?redeem=aHR0cHM6Ly8xZHJ2Lm1zL3UvcyFBaW1CZ1lWN0pqVGxnY2Q5d2k0MzJ1aXRNZ1RNLXc%5FZT1manVKYm0&cid=E534267B85818129&id=E534267B85818129%2125597&parId=E534267B85818129%2125575&o=OneUp)] uses ViTAEv2-S as its backbone and is trained on `Synth150K+Total-Text+MLT17+IC13+IC15+TextOCR`.
|
|
@@ -55,7 +55,7 @@ Please download this weight from the official [MaskTextSpotterV3](https://github
|
|
| 55 |
### `resources/data/arial_unicode_ms.ttf`
|
| 56 |
Since the Arial font cannot be redistributed, please obtain it through your operating system or another legal source. As an alternative, you may use an open font such as Liberation Sans (resources/data/LiberationSans-Regular.ttf). However, please note that we have observed a drop of 1–2 points in OCR accuracy on the Mario-Eval benchmark when using AnyText with Liberation Sans under our best configuration.
|
| 57 |
|
| 58 |
-
## Dataset details
|
| 59 |
|
| 60 |
### `weight/`
|
| 61 |
Pretrained weights used for various modules in the Type-R pipeline:
|
|
|
|
| 11 |
|
| 12 |
This repository contains model weights and data resources used in the **Type-R** project. The dataset is designed to support text-to-image generation, OCR, text erasing, editing, and evaluation pipelines used in the Type-R system.
|
| 13 |
|
| 14 |
+
## 📘 Directory structure
|
| 15 |
|
| 16 |
⚠️ The code in the [repository](https://github.com/CyberAgentAILab/Type-R) is designed to operate directly using this structure.
|
| 17 |
|
|
|
|
| 45 |
|
| 46 |
</pre>
|
| 47 |
|
| 48 |
+
## 📘 ⚠️Manual download required⚠️
|
| 49 |
### `resources/weight/ocr/solo.pth`
|
| 50 |
Please download this weight from the official [Deeosolo](https://github.com/ViTAE-Transformer/DeepSolo) implementation.
|
| 51 |
[[link](https://onedrive.live.com/?redeem=aHR0cHM6Ly8xZHJ2Lm1zL3UvcyFBaW1CZ1lWN0pqVGxnY2Q5d2k0MzJ1aXRNZ1RNLXc%5FZT1manVKYm0&cid=E534267B85818129&id=E534267B85818129%2125597&parId=E534267B85818129%2125575&o=OneUp)] uses ViTAEv2-S as its backbone and is trained on `Synth150K+Total-Text+MLT17+IC13+IC15+TextOCR`.
|
|
|
|
| 55 |
### `resources/data/arial_unicode_ms.ttf`
|
| 56 |
Since the Arial font cannot be redistributed, please obtain it through your operating system or another legal source. As an alternative, you may use an open font such as Liberation Sans (resources/data/LiberationSans-Regular.ttf). However, please note that we have observed a drop of 1–2 points in OCR accuracy on the Mario-Eval benchmark when using AnyText with Liberation Sans under our best configuration.
|
| 57 |
|
| 58 |
+
## 📘 Dataset details
|
| 59 |
|
| 60 |
### `weight/`
|
| 61 |
Pretrained weights used for various modules in the Type-R pipeline:
|