AI4Industry commited on
Commit
f40ec00
·
verified ·
1 Parent(s): 6eb011a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -2
README.md CHANGED
@@ -78,12 +78,16 @@ This repo provids the training data and evaluation data for MolParser, proposed
78
 
79
  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.
80
 
81
- * Training Dataset: More than 7.7M training data in `pretrain_synthetic_7M` subset;
 
 
82
 
83
  * Validation Dataset: 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;
84
 
85
- * WildMol-20k: 20k molecule structure images cropped from real patents or paper `test_simple_10k`(ordinary)subset and `test_markush_10k`(markush)subset;
86
 
87
  As the paper is still **under review**, this data is provided **anonymously**. More information will be provided after the paper has been accepted.
88
 
 
 
89
  [**Anonymous Demo: Click Here**](http://101.126.35.171:50008/)
 
78
 
79
  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.
80
 
81
+ * Pretraining Dataset: More than 7.7M synthetic training data in `pretrain_synthetic_7M` subset;
82
+
83
+ * SFT Dataset [WIP]: We have identified some annotation errors and are currently re-cleaning the data. Once completed, we will partially open-source it.
84
 
85
  * Validation Dataset: 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;
86
 
87
+ * WildMol-20k Benchmark: 20k molecule structure images cropped from real patents or paper `test_simple_10k`(ordinary)subset and `test_markush_10k`(markush)subset;
88
 
89
  As the paper is still **under review**, this data is provided **anonymously**. More information will be provided after the paper has been accepted.
90
 
91
+ In the future, we will continue to re-clean this dataset, open-source more data, update model checkpoints, refresh the benchmark results, and release the training code.
92
+
93
  [**Anonymous Demo: Click Here**](http://101.126.35.171:50008/)