graphstorm
graphstorm copied to clipboard
awslabs
Enterprise graph machine learning framework for billion-scale graphs for ML scientists and data scientists.
Right now for the distributed pipeline we would first create raw id mappings using GSProcessing, which the second step of DistPart does not touch. When a user runs a training...
*Issue #, if available:* *Description of changes:* This is a draft PR because I found the recall@K is extremly low and it seems the remap get wrong from reading embeddings....
With the number of arguments that we parse, it's likely that an arg can be misspelled. e.g. I ran into the case where I was trying to pass `--save-embeds-path` instead...
When running a link prediction training task, it won't give --node-emb-dir to remap script when --save-embedding-path is True.
Add doc about saving configuration files when processing graph input data (e.g. --output-conf-file). The saved configuration file will record how data are transformed (numerical features, categorical features, etc.). For example:...
*Issue #, if available:* N/A *Description of changes:* Adds PyTorch Lightning support, `LightningModule` and `LightningDataModule` for Node GNN model and also a Jupyter nodebook demonstrating how it works. By submitting...
GraphStorm should support multi-task learning as its built-in capability. * Unsupervised multi-task learning: Other than using link prediction as the sole supervision signal for unsupervised learning, one can add other...
https://github.com/awslabs/graphstorm/blob/main/python/graphstorm/gconstruct/file_io.py#L153 The default behavior of numpy.array(string) will allocate too much memory than required, which is not efficient.
I got following error using graphstorm ``` Traceback (most recent call last): File "/graphstorm/python/graphstorm/run/gsgnn_lp/gsgnn_lp.py", line 254, in main(gs_args) File "/graphstorm/python/graphstorm/run/gsgnn_lp/gsgnn_lp.py", line 106, in main train_data = GSgnnLPTrainData(config.graph_name, File "/graphstorm/python/graphstorm/dataloading/dataset.py", line...
In https://github.com/awslabs/graphstorm/pull/760, we use GPEFT module as the gnn_encoder for link prediction. We still need to modify the node classification code to use GPEFT component and correct its model save...