Closed leilayesufu closed 15 hours ago
@GemmaTuron I couldn't access the complete article.
Hi @leilayesufu I've asked for it on the slack channel, you can start working on the code meanwhile
try to run the BERT model with their pretrained checkpoints
@GemmaTuron I tried running the code as seen here https://github.com/mldlproject/2021-NPBERT-Antimalaria/tree/main/training/NPBERT_pretrained_model/save_model
I got the following error
Traceback (most recent call last): File "extract_feature.py", line 41, in <module> tokenizer=btokenizer File "/home/leila/miniconda3/envs/BERT/lib/python3.7/site-packages/transformers/pipelines/__init__.py", line 340, in pipeline framework = framework or get_framework(model) File "/home/leila/miniconda3/envs/BERT/lib/python3.7/site-packages/transformers/pipelines/base.py", line 66, in get_framework model = AutoModel.from_pretrained(model, revision=revision) File "/home/leila/miniconda3/envs/BERT/lib/python3.7/site-packages/transformers/models/auto/modeling_auto.py", line 729, in from_pretrained pretrained_model_name_or_path, *model_args, config=config, **kwargs File "/home/leila/miniconda3/envs/BERT/lib/python3.7/site-packages/transformers/modeling_utils.py", line 984, in from_pretrained pretrained_model_name_or_path, OSError: Error no file named ['pytorch_model.bin', 'tf_model.h5', 'model.ckpt.index'] found in directory ./save_model or
from_tfset to False
This suggests that the pretrained model files are not present in the save_model directory. I tried to download the files from the link provided but it's a broken link
The model is on hold since the checkpoints are not provided. I've opened an issue on their repository asking for them.
@GemmaTuron did they ever get back?
no, I'll do a follow up!
This will not be implemented since we do not have the checkpoints and the authors haven't responded.
Model Name
Predicting Antimalarial Activity in Natural Products Using Pretrained Bidirectional Encoder Representations from Transformers
Model Description
This model uses a molecular encoding scheme based on Bidirectional Encoder Representations from Transformers (BERT), employing a pretrained encoding model called NPBERT. Four machine learning algorithms—k-Nearest Neighbors (k-NN), Support Vector Machines (SVM), eXtreme Gradient Boosting (XGB), and Random Forest (RF) were employed to create prediction models. The results indicate that SVM models outperform others, and the proposed NPBERT molecular encoding scheme is more effective than existing methods.
Slug
NPBERT-Antimalaria
Tag
Malaria, P.falciparum
Publication
https://pubs.acs.org/doi/full/10.1021/acs.jcim.1c00584
Source Code
https://github.com/mldlproject/2021-NPBERT-Antimalaria
License
None