Update README.md
Browse files
README.md
CHANGED
@@ -94,13 +94,13 @@ This repo provides the training data and evaluation data for MolParser, proposed
|
|
94 |
|
95 |
MolParser-7M contains nearly 8 million paired image-SMILES data. It should be noted that the caption of image is our extended-SMILES format, which suggested in our paper.
|
96 |
|
97 |
-
* MolParser-Pretrain
|
98 |
|
99 |
-
* MolParser-SFT [WIP]
|
100 |
|
101 |
-
* MolParser-Val
|
102 |
|
103 |
-
* WildMol Benchmark
|
104 |
|
105 |
As the paper is still **under review**, this data is provided **anonymously**. More information will be provided after the paper has been accepted.
|
106 |
|
|
|
94 |
|
95 |
MolParser-7M contains nearly 8 million paired image-SMILES data. It should be noted that the caption of image is our extended-SMILES format, which suggested in our paper.
|
96 |
|
97 |
+
* **MolParser-Pretrain**: More than 7.7M synthetic training data in `pretrain_synthetic_7M` subset;
|
98 |
|
99 |
+
* **MolParser-SFT [WIP]**: Nearly 400k samples for fine-tuning stage in `sft_real` subset. (We are organizing an OCSR competition based on MolParser-7M, so we have reserved part of the MolParser-SFT data for the competition. Stay tuned!)
|
100 |
|
101 |
+
* **MolParser-Val**: A small validation set carefully selected in-the-wild in `valid` subset. It can be used to quickly valid the model ability during the training process;
|
102 |
|
103 |
+
* **WildMol Benchmark**: 20k molecule structure images cropped from real patents or paper, `test_simple_10k`(WildMol-10k)subset and `test_markush_10k`(WildMol-10k-M)subset;
|
104 |
|
105 |
As the paper is still **under review**, this data is provided **anonymously**. More information will be provided after the paper has been accepted.
|
106 |
|