Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
task_categories:
|
| 4 |
+
- feature-extraction
|
| 5 |
+
---
|
| 6 |
+
# MechLLM
|
| 7 |
+
This is a synthetic dataset generated for electron movement extraction.
|
| 8 |
+
|
| 9 |
+
## Data Description
|
| 10 |
+
The `synthetic` folder contains the synthetic reaction image from the pistachio dataset with curved arrow programatically drawn on the image.
|
| 11 |
+
|
| 12 |
+
The `number` folder contains the related number labels of the atoms.
|
| 13 |
+
|
| 14 |
+
the json file for corresponding image folders contains the start and end atom of electron movements, and the number of electron movements.
|
| 15 |
+
<pre><code>```
|
| 16 |
+
.
|
| 17 |
+
├── number/
|
| 18 |
+
│ ├── generated_images_debug/
|
| 19 |
+
│ ├── generated_images_debug_multiline/
|
| 20 |
+
│ ├── generated_images_debug_multiline2/
|
| 21 |
+
│ ├── generated_images_debug_multiline_arrow/
|
| 22 |
+
│ ├── generated_images_debug_single_multire/
|
| 23 |
+
│ └── images/
|
| 24 |
+
├── synthetic/
|
| 25 |
+
│ ├── generated_images_debug/
|
| 26 |
+
│ ├── generated_images_debug_multiline/
|
| 27 |
+
│ ├── generated_images_debug_multiline2/
|
| 28 |
+
│ ├── generated_images_debug_multiline_arrow/
|
| 29 |
+
│ ├── generated_images_debug_single_multire/
|
| 30 |
+
│ └── images/
|
| 31 |
+
├── generated_images_debug.json
|
| 32 |
+
├── generated_images_debug_multiline.json
|
| 33 |
+
├── generated_images_debug_multiline2.json
|
| 34 |
+
├── generated_images_debug_multiline_arrow.json
|
| 35 |
+
├── generated_images_debug_single_multire.json
|
| 36 |
+
└── images.json
|
| 37 |
+
```</code></pre>
|
| 38 |
+
## Dataset Utilization
|
| 39 |
+
The dataset aims to benefit curved arrow information extraction for understanding chemical reaction mechanisms.
|
| 40 |
+
|
| 41 |
+
This dataset can be used in different models, benchmarking the performance of object recognition.
|