Update README.md

README.md

@@ -3,90 +3,88 @@ license: apache-2.0
 tags:
 - biology
 - chemistry
+configs:
+  - config_name: bace
+    data_files:
+      - split: train
+        path: bace/train.csv
+      - split: test
+        path: bace/test.csv
+      - split: val
+        path: bace/valid.csv
+  - config_name: bbbp
+    data_files:
+      - split: train
+        path: bbbp/train.csv
+      - split: test
+        path: bbbp/test.csv
+      - split: val
+        path: bbbp/valid.csv
+  - config_name: clintox
+    data_files:
+      - split: train
+        path: clintox/train.csv
+      - split: test
+        path: clintox/test.csv
+      - split: val
+        path: clintox/valid.csv
+  - config_name: esol
+    data_files:
+      - split: train
+        path: esol/train.csv
+      - split: test
+        path: esol/test.csv
+      - split: val
+        path: esol/valid.csv
+  - config_name: freesolv
+    data_files:
+      - split: train
+        path: freesolv/train.csv
+      - split: test
+        path: freesolv/test.csv
+      - split: val
+        path: freesolv/valid.csv
+  - config_name: hiv
+    data_files:
+      - split: train
+        path: hiv/train.csv
+      - split: test
+        path: hiv/test.csv
+      - split: val
+        path: hiv/valid.csv
+  - config_name: lipo
+    data_files:
+      - split: train
+        path: lipo/train.csv
+      - split: test
+        path: lipo/test.csv
+      - split: val
+        path: lipo/valid.csv
+  - config_name: qm9
+    data_files:
+      - split: train
+        path: qm9/train.csv
+      - split: test
+        path: qm9/test.csv
+      - split: val
+        path: qm9/valid.csv
+  - config_name: sider
+    data_files:
+      - split: train
+        path: sider/train.csv
+      - split: test
+        path: sider/test.csv
+      - split: val
+        path: sider/valid.csv
+  - config_name: tox21
+    data_files:
+      - split: train
+        path: tox21/train.csv
+      - split: test
+        path: tox21/test.csv
+      - split: val
+        path: tox21/valid.csv
 ---
 
-# MoLFormer
+# MoleculeNet Benchmark
 
-**MoLFormer** is a large-scale chemical language model designed with the intention of learning a model trained on small molecules which are represented as SMILES strings. MoLFormer leverges Masked Language Modeling and employs a linear attention Transformer combined with rotary embeddings.
-
-![MoLFormer](https://media.github.ibm.com/user/4935/files/594363e6-497b-4b91-9493-36ed46f623a2)
-
-An overview of the MoLFormer pipeline is seen in the image above. One can see that the transformer based neural network model is trained on a large collection of chemical molecules represented by SMILES sequences from two public chemical datasets PubChem and Zinc in a self-supervised fashion.  The MOLFORMER architecture was designed with an efficient linear attention mechanism and relative positional embeddings with the goal of learning a meaningful and compressed representation of chemical molecules. After training the MOLFORMER foundation model was then adopted to different downstream molecular property prediction tasks via fine-tuning on task-specific data. To further test the representative power of MOLFORMER the MOLFORMER encodings were used to recover molecular similarity, and analysis on the correspondence between the interatomic spatial distance and attention value for a given molecule was performed.
-
-## Finetuning Datasets
-Just as with the pretraining data the code expects the finetuning datasets to be in the following hierarchy. These datasets were provided in the finetune_datasets.zip
-
-```
-data/
-├── bace
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── bbbp
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── clintox
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── esol
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── freesolv
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── hiv
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-├── lipo
-│   ├── lipo_test.csv
-│   ├── lipo_train.csv
-│   └── lipo_valid.csv
-├── qm9
-│   ├── qm9.csv
-│   ├── qm9_test.csv
-│   ├── qm9_train.csv
-│   └── qm9_valid.csv
-├── sider
-│   ├── test.csv
-│   ├── train.csv
-│   └── valid.csv
-└── tox21
-    ├── test.csv
-    ├── tox21.csv
-    ├── train.csv
-    └── valid.csv
-```
-
-
-## Citations
-```
-@article{10.1038/s42256-022-00580-7,
-year = {2022},
-title = {{Large-scale chemical language representations capture molecular structure and properties}},
-author = {Ross, Jerret and Belgodere, Brian and Chenthamarakshan, Vijil and Padhi, Inkit and Mroueh, Youssef and Das, Payel},
-journal = {Nature Machine Intelligence},
-doi = {10.1038/s42256-022-00580-7},
-abstract = {{Models based on machine learning can enable accurate and fast molecular property predictions, which is of interest in drug discovery and material design. Various supervised machine learning models have demonstrated promising performance, but the vast chemical space and the limited availability of property labels make supervised learning challenging. Recently, unsupervised transformer-based language models pretrained on a large unlabelled corpus have produced state-of-the-art results in many downstream natural language processing tasks. Inspired by this development, we present molecular embeddings obtained by training an efficient transformer encoder model, MoLFormer, which uses rotary positional embeddings. This model employs a linear attention mechanism, coupled with highly distributed training, on SMILES sequences of 1.1 billion unlabelled molecules from the PubChem and ZINC datasets. We show that the learned molecular representation outperforms existing baselines, including supervised and self-supervised graph neural networks and language models, on several downstream tasks from ten benchmark datasets. They perform competitively on two others. Further analyses, specifically through the lens of attention, demonstrate that MoLFormer trained on chemical SMILES indeed learns the spatial relationships between atoms within a molecule. These results provide encouraging evidence that large-scale molecular language models can capture sufficient chemical and structural information to predict various distinct molecular properties, including quantum-chemical properties. Large language models have recently emerged with extraordinary capabilities, and these methods can be applied to model other kinds of sequence, such as string representations of molecules. Ross and colleagues have created a transformer-based model, trained on a large dataset of molecules, which provides good results on property prediction tasks.}},
-pages = {1256--1264},
-number = {12},
-volume = {4}
-}
-```
-
-```
-@misc{https://doi.org/10.48550/arxiv.2106.09553,
-  doi = {10.48550/ARXIV.2106.09553},
-  url = {https://arxiv.org/abs/2106.09553},
-  author = {Ross, Jerret and Belgodere, Brian and Chenthamarakshan, Vijil and Padhi, Inkit and Mroueh, Youssef and Das, Payel},
-  keywords = {Machine Learning (cs.LG), Computation and Language (cs.CL), Biomolecules (q-bio.BM), FOS: Computer and information sciences, FOS: Computer and information sciences, FOS: Biological sciences, FOS: Biological sciences},
-  title = {Large-Scale Chemical Language Representations Capture Molecular Structure and Properties},
-  publisher = {arXiv},
-  year = {2021},
-  copyright = {arXiv.org perpetual, non-exclusive license}
-}
-```