graphstorm
graphstorm copied to clipboard
awslabs
Enterprise graph machine learning framework for billion-scale graphs for ML scientists and data scientists.
*Issue #, if available:* *Description of changes:* By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
Got this when running with a custom DGL installation. Using a training command e.g. ```bash python3 -m graphstorm.run.gs_node_classification \ --extra-envs LD_LIBRARY_PATH="/opt/gs-venv/lib/python3.9/site-packages/dgl/:$LD_LIBRARY_PATH" \ --num-trainers 1 \ --num-servers 1 \ --num-samplers 0...
Error code is like: ``` self._sparse_embeds[ntype] = DistEmbedding(g.number_of_nodes(ntype), self.embed_size, embed_name + '_' + ntype, init_emb, part_policy=part_policy) ...... if len(input_nodes[ntype]) == 0: dtype = self.sparse_embeds[ntype].weight.dtype embs[ntype] = th.zeros((0, self.sparse_embeds[ntype].embedding_dim), device=device, dtype=dtype)...
The V0.3 release will focus on optimizing the code structure and adding new features such as multi-task learning. It will also includes other improvements and features. - [Improvement] Graph data...
*Issue #, if available:* *Description of changes:* This PR add a new Jupyter notebook example for demonstrating link prediction pipeline. For the notebook, this PR also adds a simple RGCN...
*Issue #, if available:* #835 *Description of changes:* By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your...
*Issue #, if available:* #789 *Description of changes:* ### Graph construction Update GraphStorm input config parsing to support multi-task learning. Allow user to specify to specify multiple training tasks for...
Originally, we set remove_target_edge_type to True. So in edge prediction tasks (classification and regression) the target edge type is not included in message passing by default. But, is this the...
*Issue #, if available:* *Description of changes:* * We change the way we configure the Spark env, to only create our own config for SageMaker, as EMR/EMRS will have pre-configured...