Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Bharath Ramsundar
    @rbharath
    Running 10 minutes for OH today. Will be on at 9:10am PST
    Bharath Ramsundar
    @rbharath
    I won't be able to make OH today but will be back on tomorrow as usual
    Bharath Ramsundar
    @rbharath
    Running 15 minutes lates for OH today
    Bharath Ramsundar
    @rbharath
    Sorry forgot to mention but I am out of office today. Will be back for OH tomorrow
    hemogoblin
    @hemogoblin:anontier.nl
    [m]
    Hello world!
    Balaji
    @BalajiAI_twitter
    Is there a OH today?
    I'm waiting for someone to let me in. Time - Thursday 7.30 PM PST. Friday 9.00 AM IST
    Bharath Ramsundar
    @rbharath
    Yes there was OH today
    OH is at 7:00 pm PST, the developer call is at 7:30 pm PST
    I think you may have gotten the calls mixed up, sorry!
    There will be OH tomorrow at 9:00am PST / 10:30pm IST
    Can you make it then?
    Balaji
    @BalajiAI_twitter
    No problem. yes the OH timeline is quite confusing to me.
    Ok, I'll join tomorrow then.
    Bharath Ramsundar
    @rbharath
    I’m running a bit late for the deepchem dev call today so may be 15 minutes late
    Balaji
    @BalajiAI_twitter
    @rbharath I'm confused again. Since you're in PST time zone, you said to join tomorrow. But in ist, it is still today.
    Ok. I'll mail you to add me to OH hours meeting. Otherwise I'll still be confused.
    Bharath Ramsundar
    @rbharath
    Sounds good! It can get a little confusing across time zones for sure
    Balaji
    @BalajiAI_twitter
    Ah, yes!
    Bharath Ramsundar
    @rbharath
    Running 5minutes late to OH
    Balaji
    @BalajiAI_twitter
    I've been trying to install deepchem from source. But it shows MS Visual C++ is needed. But I aldready installed - MS Visual C++ Redistributable. Both are same right?
    Terminal 11_16_2022 1_46_04 AM.png
    Are there anyone who encountered the same?
    Balaji
    @BalajiAI_twitter
    Bharath Ramsundar
    @rbharath
    Running 10 minutes late to OH today
    Balaji
    @BalajiAI_twitter
    ok bharath. I'll join today.
    Bharath Ramsundar
    @rbharath
    Sorry have a scheduling conflict so I don’t think I can make OH today. Back as usual tomorrow
    Bharath Ramsundar
    @rbharath
    Still out so won't be at OH / dev call today. Back as usual tomorrow
    Bharath Ramsundar
    @rbharath
    Running a few minutes late
    Bharath Ramsundar
    @rbharath
    No office hours or meetings for the rest of this week for Thanksgiving. Happy Thanksgiving to everyone!
    fnthibaud
    @fnthibaud
    Hi all, I'm new to deepchem so I have a noob question. I wanted to train a deep learning model on the QM9 dataset so I started with a DTNN model with the same hyperparamters as in the qm9_DTNN.py example. This got me very poor pearson r2 scores, but I saw in issue #2475 on github that this might be a bug related to that type of model. However, after trying to train a GraphConv model on it this only gave me slightly better results (~0.33 pearson_r2 after 100 epochs). I used the default hyperparameters for the GraphConv model. Does anyone know if this can be resolved by hyperparameter tuning and if so, can you give me some advice on how to do this?
    Bharath Ramsundar
    @rbharath
    Sorry for the late notice, but I just realized I can’t make OH today. Will be back as usual tomorrow
    Bharath Ramsundar
    @rbharath
    @fnthibaud Try using simpler models (say RF + Coulomb fingerprints) to start and see how that does
    Bharath Ramsundar
    @rbharath
    Running 10 minutes late to OH
    Bharath Ramsundar
    @rbharath
    My apologies for the late notice but I can't make oh today. Will be back as usual tomorrow!
    Cretaceous
    @cretaceousmart
    Hi everyone. I wonder if there is an API in Deepchem that allows me to obtain a fingerprint of a drug structure?
    Since I want to create a new dataset that is outside of deepchem :)
    Bharath Ramsundar
    @rbharath
    @cretaceousmart There isn't an API but you can write a script to fingerprint compounds yourself and set up a server
    Cretaceous
    @cretaceousmart
    Many Thanks :)
    Bharath Ramsundar
    @rbharath
    I won't be at OH today because of a conflict. Back tomorrow!
    br3g19
    @br3g19
    Hi, is there a way to have Keras model that takes in 2 inputs and runs it through 2 parallel set of layers that have graph conv and graph pool layers in the set of layers with the outputs of the 2 sets being combined with layers.Add layer then a dense layer after that then output. All the examples of custom graph convolution models seem to be made for only one input set
    Bharath Ramsundar
    @rbharath
    @br3g19 You'd have to make a custom model that does that. You can probably reuse some deepchem layers to help though
    1 reply
    br3g19
    @br3g19
    @rbharath I am trying to do that but every example of using the graph layers with deepchem for a custom model used one input variable and I tried making the call function input to take in multiple inputs like in the way keras.model would be set up to take multiple inputs however function didn't accept that. Then tried to build the keras.model by calling the inputs and output layers using the keras.model and set 2 inputs and an output for the model however it seems that the output of the graphconv layer isn't accepted by either the ker.batch normilisation layer (unknown shape error) or dc.graph pool layer (didn't accept data type from the graph conv I think)
    Heta Gandhi
    @gandhi_heta_twitter
    @rbharath Are deepchem featurization methods parallelizable? I found this issue that was closed in 2016 deepchem/deepchem#267. Does the user still bear responsibility to parallelize featurization using multiple dataloaders? Or is there an argument/feature for user to specify if featurization should be parallelized?
    Bharath Ramsundar
    @rbharath
    @gandhi_heta_twitter Good question. I think the featurization methods are safely parallelizable. The responsibility is on the user, but this should hopefully be a simple multiprocessing script on the user side
    We used to support multiprocessing in deepchem, but it was brittle and had to make assumptions on the user compute setup. Having users do it lowered bugs but the tradeoff is it's a bit more effort from the user. I think one of the tutorials gives an example
    The large scale virtual screening tutorial
    Heta Gandhi
    @gandhi_heta_twitter
    @rbharath Thanks for the clarification! Can you paste a link to that tutorial you mentioned? That would be very helpful.
    This has a multiprocessing example script
    Bharath Ramsundar
    @rbharath
    DeepChem office hours and meetings are paused from now till January 3rd for the holidays. We will resume regular meetings then. I hope everyone has a wonderful holiday break!