nicola-decao / molgan Goto Github PK

If activation is None, it seems to return output rather than activation. This problem is also in line 24 of the same file. please check.

Thank you

Hello. I am a Japanese student.
Sorry if there is an inappropriate expression.

I have 2 questions.

• How can I output the generated results?
• What is different from paper experiments?

Thank you

Hello. I am a student at Korea university.

I try to implement MolGAN with Pytorch. Your implementation is very helpful. Thank you.
In the meantime, I have a question. I referenced the article and the previous Graph Net, but I still do not know what 'la' is.
'la' plays an important role in the model, what is 'la'?

Answers I'll wait. Thank you.

Thanks for your great work.

While I have a question about the bond type set, in your paper you say the bond_type Y is set to 4 (single, double, triple, and no bond).

However, the default result of function

may contain 5 bond types (and also aromatic bonds)

I wonder if there is any method to abandon the aromatic type or am I just missing something?

Hello!

Thank you for the implementation of MolGAN

How should I update your code to use with bigger molecules (~20-30 heavy atoms)?

It seems, that only passing another dataset (with bigger molecules) after 100 epochs model generated very bad SMILES like these:

*.*.*.*.*CC.C.C.C.C.C.C.C.C.C.C.C.CC.CCN.CCNC.O.O.O
*.*.*.*.*.*.*.*.*C(*)C.C.C.C.C.C.C.CC.CCN.CCO.CO.N.N
*.*.*.*.*.*.*.*.*.*=C(C)CC.C.C.C.C.C.C.CC.CCC.CCCCO
*.*.*.*.*.*C.*C*.C.C.C.C.C.C.C.C=C.CC.CC(C)F.CNC.N.O
*.*.*.*.*.*.*.*.*.*C.*CC.C.C.C.C.C.C.C.C.CCNCO.CN.O
*.*.*.*.*.*.*.*.*.*OCC.C.C.C.CCC.CCCC.CSC.N.N.O.O
*.*.*.*.*.*.*.*.*.*.*.*.*CCOC(C)C*.*CN=*.C.C.C.CC.O
*.*.*.*.*.*.*CC.C.C.C.C.C.C.C.C.C.C.C=CC.CC.CC(C)S.N.N
*.*.*.*.*.*.*.*.*C.*C.C.C.C.C.C.C*CC.CC.CC.COO.N.O
*.*.*.*.*.*.*.*C(*)(*)O.*CC(C)CC.C.C.C.CCC.CCCN.N.O
*.*.*.*.*.*.*.*.*.*.*.*.*CN.*O.C.C.C.C.CCC.CCCC.O.S
*.*.*.*.*.*.*.*.*.*C.*C.*CSCC(C)C(CC)C(C)*C(C)C.C.C
*.*.*.*.*.*.*CC(C)CCC.*N.C.C.C.C.C.C.C.CN.CO.F.O.O.O
*.*.*.*.*.*.*.*.*C.*C.*C.*C(C)C.C.C.C.C.C.C.CC.CCN.O
*.*.*.*.*.*.*CC.*NC.C.C.C.C.C.C.C.C.C.CC.N.N.NCC(O)O
*.*C.C.C.C.C.C.C.C.C.C.C.C.C.C.C.CC.CNCO.N.N*C(=O)O.O
*.*.*.*.*.**.*C.*CC=[SH]*.C.C.C.C.C.C.C.CC.CC.CCC.N.O
*.*.*.*.*.*.*.*C(C)(F)CC(C)CC.C.C.C.C.C.CC.CCC.N.N.N.O

I meet errors when I convert the data format using sparse_molecular_dataset.py, like this:

[02:21:06] Explicit valence for atom # 1 C, 5, is greater than permitted
[02:21:06] ERROR: Could not sanitize molecule ending on line 9097
[02:21:06] ERROR: Explicit valence for atom # 1 C, 5, is greater than permitted
[02:21:06] Explicit valence for atom # 1 C, 5, is greater than permitted
[02:21:06] ERROR: Could not sanitize molecule ending on line 35785
[02:21:06] ERROR: Explicit valence for atom # 1 C, 5, is greater than permitted
[02:21:06] Explicit valence for atom # 4 C, 5, is greater than permitted
[02:21:06] ERROR: Could not sanitize molecule ending on line 62866
[02:21:06] ERROR: Explicit valence for atom # 4 C, 5, is greater than permitted
[02:21:06] Explicit valence for atom # 2 C, 5, is greater than permitted
[02:21:06] ERROR: Could not sanitize molecule ending on line 66832

and finally, the error message is

Traceback (most recent call last):
  File "sparse_molecular_dataset.py", line 326, in <module>
    data.generate('data/gdb9.sdf', filters=lambda x: x.GetNumAtoms() <= 9)
  File "sparse_molecular_dataset.py", line 53, in generate
    self._generate_AX()
  File "sparse_molecular_dataset.py", line 127, in _generate_AX
    pr = ProgressBar(60, len(self.data))
  File "/home/zeyuwen/workspace/anaconda3/envs/tf/lib/python3.6/site-packages/progress_bar.py", line 113, in __init__
    self.title = title + ": "
TypeError: unsupported operand type(s) for +: 'int' and 'str'

Can you give me some advice?

Hi!
I have an error as below when I run exmaple.py.

'QED score': 0.5195647414042559,
'SA score': 0.7086636282519222,
'diversity score': 0.6641403755404658,
'drugcandidate score': 0.5692159133414297,
'la': 1.0,
'logP score': 0.3386603861602298,
'loss D': -66.3097,
'loss G': 32.144226,
'loss RL': -0.40582168,
'loss V': 0.2708291,
'novel score': 100.0,
'unique score': 0.5204460966542751,
'valid score': 80.69999814033508}
Traceback (most recent call last):
File "example.py", line 214, in
_test_update=_test_update)
File "/Users/owner/MolGAN/utils/trainer.py", line 170, in train
_test_step(self.model, self.optimizer, batch_dim, eval_batch, start_time, _test_update)
File "/Users/owner/MolGAN/utils/trainer.py", line 103, in _test_step
self.load(directory)
File "/Users/owner/MolGAN/utils/trainer.py", line 39, in load
saver.restore(self.session, '{}/{}.ckpt'.format(directory, 'model'))
File "/anaconda3/lib/python3.6/site-packages/tensorflow/python/training/saver.py", line 1538, in restore
+ compat.as_text(save_path))
ValueError: The passed save_path is not a valid checkpoint: //model.ckpt

Please let me know how to start example.py more details.

Hello,

It is nice to see the codes for the very interesting paper. I was wondering which part of the code I should use to try to reproduce results from Table 3 (Comparison with different algorithms on QM9).

Thanks in advance.