[Model & Dataset] facebook & sp2gcl #201
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dddg617
reviewed
May 20, 2024
gammagl/datasets/facebook.py
Outdated
| x = tlx.convert_to_tensor(data['features'], dtype=tlx.float32) | ||
| y = tlx.convert_to_tensor(data['target'], dtype=tlx.int64) | ||
| edge_index = tlx.convert_to_tensor(data['edges'], dtype=tlx.int64) | ||
| edge_index = edge_index.T |
There was a problem hiding this comment.
Have you tried if this can work in the other backend like 'mindspore'?
gammagl/datasets/facebook.py
Outdated
|
|
||
| def __init__( | ||
| self, | ||
| root: str, |
There was a problem hiding this comment.
Currently, this argument can be optional, as we have a cached mechanism compared to PyG.
dddg617
reviewed
May 27, 2024
examples/sp2_gcl/node_main.py
Outdated
| loss = 0.5 * tlx.losses.softmax_cross_entropy_with_logits(logits, labels) + 0.5 * tlx.losses.softmax_cross_entropy_with_logits(logits.transpose(-2, -1), labels) | ||
| return loss | ||
| def main(args): | ||
| global edge, e, u, test_idx |
examples/sp2_gcl/node_main.py
Outdated
| val_idx = tlx.where(data.val_mask)[0] | ||
| test_idx = tlx.where(data.test_mask)[0] | ||
| else: | ||
| train_idx, val_idx, test_idx = split(y) |
There was a problem hiding this comment.
I do not think this is useful. Usually, the train, valid, test split should be done in the dataset. You may directly use data.train_mask .etc to get the idx instead of add a new function in the util.
gammagl/models/sp2gcl.py
Outdated
| import tensorlayerx as tlx | ||
| import tensorlayerx.nn as nn | ||
| from gammagl.layers.conv import GCNConv | ||
| class Encoder(nn.Module): |
There was a problem hiding this comment.
I remember there are many Encoder in the gammagl. I do not recommend you to place this function here.
gammagl/models/sp2gcl.py
Outdated
| return x | ||
|
|
||
|
|
||
| class MLP(nn.Module): |
|
|
||
|
|
||
|
|
||
| class EigenMLP(nn.Module): |
gammagl/utils/split.py
Outdated
| import tensorlayerx as tlx | ||
| import numpy as np | ||
|
|
||
| def split(y): |
There was a problem hiding this comment.
This util is useless, remove it.
There was a problem hiding this comment.
from sklearn.model_selection import train_test_split
gyzhou2000
reviewed
Jun 3, 2024
examples/sp2_gcl/readme.md
Outdated
| @ -0,0 +1,40 @@ | ||
| # Graph Contrastive Learning with Stable and Scalable | ||
|
|
||
| - Paper link: [https://proceedings.neurips.cc/paper_files/paper/2023/file/8e9a6582caa59fda0302349702965171-Paper-Conference.pdf](https://arxiv.org/abs/2201.11349) |
gammagl/models/sp2gcl.py
Outdated
Comment on lines
39
to
42
| period_e = e.unsqueeze(1) * tlx.pow(2, period_term) | ||
| # period_e = period_e.to(u.device) | ||
| fourier_e = tlx.concat([tlx.sin(period_e), tlx.cos(period_e)], axis=-1) | ||
| h = u @ fourier_e |
There was a problem hiding this comment.
tlx.unsqueeze，矩阵乘用tlx.matmul
examples/sp2_gcl/readme.md
Outdated
Comment on lines
32
to
34
| | PubMed | 82.3±0.3 | OOM | | ||
| | Wiki-CS | 79.42±0.19 | 76.79 ± 0.61 | | ||
| | Facebook | 90.43±0.13 | 85.35±0.26 | |
gammagl/models/sp2gcl.py
Outdated
Comment on lines
29
to
30
| self.phi = nn.Sequential(nn.Linear(1, 16), nn.ReLU(), nn.Linear(16, 16)) | ||
| self.psi = nn.Sequential(nn.Linear(16, 16), nn.ReLU(), nn.Linear(16, 1)) |
tests/datasets/test_facebook.py
Outdated
Comment on lines
5
to
7
| dataset = FacebookPagePage(root='data/facebook') | ||
| g = dataset[0] | ||
| pass |
There was a problem hiding this comment.
节点数量，边数量，节点特征维度，类别数量都判断一下
examples/sp2_gcl/sp2gcl_trainer.py
Outdated
Comment on lines
133
to
134
| parser.add_argument('--seed', type=int, default=0) | ||
| parser.add_argument('--cuda', type=int, default=3) |
There was a problem hiding this comment.
seed去点，设置device参考其他trainer写法
examples/sp2_gcl/sp2gcl_trainer.py
Outdated
Comment on lines
42
to
58
| def compute_laplacian(data): | ||
|
|
||
| edge_index = data.edge_index | ||
| num_nodes = data.num_nodes | ||
| row, col = edge_index | ||
| data_adj = csr_matrix((np.ones(len(row)), (row, col)), shape=(num_nodes, num_nodes)) | ||
| degree = np.array(data_adj.sum(axis=1)).flatten() | ||
| deg_inv_sqrt = 1.0 / np.sqrt(degree) | ||
| deg_inv_sqrt[np.isinf(deg_inv_sqrt)] = 0 | ||
| I = csr_matrix(np.eye(num_nodes)) | ||
| D_inv_sqrt = csr_matrix((deg_inv_sqrt, (np.arange(num_nodes), np.arange(num_nodes)))) | ||
| L = I - D_inv_sqrt.dot(data_adj).dot(D_inv_sqrt) | ||
| e, u = scipy.sparse.linalg.eigsh(L, k=100, which='SM', tol=1e-3) | ||
| data.e = tlx.convert_to_tensor(e, dtype=tlx.float32) | ||
| data.u = tlx.convert_to_tensor(u, dtype=tlx.float32) | ||
|
|
||
| return data |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Checklist
Please feel free to remove inapplicable items for your PR.
or have been fixed to be compatible with this change
Changes