Comments (4)
I'm sorry you're running into this problem. Is it possible that the second (or later) column in your csv has smile strings? By default, the the train script expects the first column in the csv to have smile strings and all later columns to have numeric labels.
from chemprop.
Happy to help. By custom regressors do you mean custom input features? If so, ChemProp does support them. You can find more information here: https://github.com/chemprop/chemprop#custom-features. Please let us know if you have any other issues!
from chemprop.
hi get error when trying to train the model
Fold 0
{'activation': 'ReLU',
'atom_messages': False,
'batch_size': 50,
'bias': False,
'checkpoint_dir': None,
'checkpoint_path': None,
'checkpoint_paths': None,
'config_path': None,
'crossval_index_dir': None,
'crossval_index_file': None,
'cuda': True,
'data_path': 'Atbvsgarlic.csv',
'dataset_type': 'classification',
'depth': 3,
'dropout': 0.0,
'ensemble_size': 1,
'epochs': 30,
'features_generator': None,
'features_only': False,
'features_path': None,
'features_scaling': True,
'ffn_hidden_size': 300,
'ffn_num_layers': 2,
'final_lr': 0.0001,
'folds_file': None,
'gpu': None,
'hidden_size': 300,
'init_lr': 0.0001,
'log_frequency': 10,
'max_data_size': None,
'max_lr': 0.001,
'metric': 'auc',
'minimize_score': False,
'multiclass_num_classes': 3,
'no_cache': False,
'num_folds': 1,
'num_lrs': 1,
'quiet': False,
'save_dir': 'tox21_checkpoints/fold_0',
'save_smiles_splits': False,
'seed': 0,
'separate_test_features_path': None,
'separate_test_path': None,
'separate_val_features_path': None,
'separate_val_path': None,
'show_individual_scores': False,
'split_sizes': [0.8, 0.1, 0.1],
'split_type': 'random',
'test': False,
'test_fold_index': None,
'undirected': False,
'use_compound_names': False,
'use_input_features': None,
'val_fold_index': None,
'warmup_epochs': 2.0}
Loading data
0% 0/12 [00:00<?, ?it/s][19:50:15] SMILES Parse Error: syntax error while parsing: Ajoene
[19:50:15] SMILES Parse Error: Failed parsing SMILES 'Ajoene' for input: 'Ajoene'
Traceback (most recent call last):
File "chemprop/train.py", line 11, in
cross_validate(args, logger)
File "/content/chemprop/chemprop/train/cross_validate.py", line 29, in cross_validate
model_scores = run_training(args, logger)
File "/content/chemprop/chemprop/train/run_training.py", line 49, in run_training
data = get_data(path=args.data_path, args=args, logger=logger)
File "/content/chemprop/chemprop/data/utils.py", line 151, in get_data
) for i, line in tqdm(enumerate(lines), total=len(lines))
File "/content/chemprop/chemprop/data/utils.py", line 151, in
) for i, line in tqdm(enumerate(lines), total=len(lines))
File "/content/chemprop/chemprop/data/data.py", line 66, in init
self.targets = [float(x) if x != '' else None for x in line[1:]]
File "/content/chemprop/chemprop/data/data.py", line 66, in
self.targets = [float(x) if x != '' else None for x in line[1:]]
ValueError: could not convert string to float: 'C=CCSSC=CCS(=O)CC=C'
0% 0/12 [00:00<?, ?it/s]
from chemprop.
thanks, awesome job lhirschfeld is it possible to train the model on customized regressors (descriptors)
do we need to overwrite configuration scripts?
from chemprop.
Related Issues (20)
- [v2 FEATURE]: Automatic device handling for loading trained models HOT 1
- transfer learning with Chemprop HOT 3
- [v2 BUG]: "RuntimeError: Detected the following values in `target`" during binary classification HOT 4
- [v2 BUG]: HOT 1
- [v2 BUG]: NaN target values set to 0.0 rather than being masked. HOT 1
- [v2 QUESTION]: Getting ValueError on classification prediction HOT 3
- [v2 BUG]: `TypeError` when manually specifying `--loss-function` as its default HOT 2
- [TODO]: Problems with infinite training HOT 2
- [BUG] fix noam scheduler definition HOT 12
- [v2 QUESTION]: Classification model prediction output. HOT 1
- [v2 QUESTION]: Is cross validation and num folds currently not complete in chemprop v2? HOT 6
- question about stereochemistry in the smiles HOT 3
- [v2 FEATURE]: Convert v0 model to v2
- [v2 FEATURE]: saving probability or the label of the predicted class as output for multiclass classification? HOT 2
- [v2 FEATURE]: allow multiple replicates for model training
- [v2 FEATURE]: allow hpopt to work on multiple splits and ensemble models
- [v2 FEATURE]: using logging to handle logging messages in all the modules
- [v2 BUG]: Ray Tune generating temporary files in different directories during hyperparameter optimization
- [v2 BUG]: MveFFN and EvidentialFFN do not work with Metrics/Loss fns HOT 5
- [v2 FEATURE]: Making it optional to save all the trials/models/checkpoints from hyperparameter optimization
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from chemprop.