-
Notifications
You must be signed in to change notification settings - Fork 41
Dev pnn pr #340
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Dev pnn pr #340
Conversation
| pnn_train_eval.py \ | ||
| --data_dir $DATA_DIR \ | ||
| --persistent_path $PERSISTENT_PATH \ | ||
| --table_size_array "43, 98, 121, 41, 219, 112, 79, 68, 91, 5, 26, 36, 70, 1447, 554, 157461, 117683, 305, 17, 11878, 629, 4, 39504, 5128, 156729, 3175, 27, 11070, 149083, 11, 4542, 1996, 4, 154737, 17, 16, 52989, 81, 40882" \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
replace with new list
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
RecommenderSystems/pnn/train_pnn.sh
Outdated
| pnn_train_eval.py \ | ||
| --data_dir $DATA_DIR \ | ||
| --persistent_path $PERSISTENT_PATH \ | ||
| --table_size_array "43, 98, 121, 41, 219, 112, 79, 68, 91, 5, 26, 36, 70, 1447, 554, 157461, 117683, 305, 17, 11878, 629, 4, 39504, 5128, 156729, 3175, 27, 11070, 149083, 11, 4542, 1996, 4, 154737, 17, 16, 52989, 81, 40882" \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
replace with new list
| return self.one_embedding.forward(ids) | ||
|
|
||
|
|
||
| class DenseLayer(nn.Module): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ueless
| one_embedding_store_type="cached_host_mem", | ||
| cache_memory_budget_mb=8192, | ||
| interaction_type="dot", | ||
| interaction_itself=False, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
uesless
| self.input_dim = embedding_vec_size * self.fields | ||
| if self.use_inner: | ||
| self.input_dim += sum(range(self.fields)) | ||
| self.inner_product_layer = InnerProductLayer( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
similar with
https://github.com/Oneflow-Inc/models/blob/main/RecommenderSystems/dlrm/dlrm_train_eval.py#L236
output_concat = None
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
OutterProductLayer not work!!!
| I = self.inner_product_layer(E) | ||
| if self.use_outter: | ||
| O = self.outter_product_layer( | ||
| E.reshape(self.fields, -1, 1, self.embedding_vec_size) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
what is the purpose of reshape here???
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
E的形状本来是(b, fields, embedding_vec_size), reshape之后是(fields, b, 1, embedding_vec_size)。为了传入opnn网络进行训练。
ShawnXuan
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
suggest to keep IPNN only currently.
No description provided.