-
Notifications
You must be signed in to change notification settings - Fork 214
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
Add more models and datasets from external packages. #42
Conversation
rayrayraykk
commented
Apr 27, 2022
•
edited
Loading
edited
- Add pre-trained transformers to NLP models.
- Add datasets from hugging face.
- add datasets from openml.
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.
LGTM. Please help us review the usage of Transformer and the datasets (e.g., SST-2)? @yxdyc Thanks!
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.
Good job, plz see the inline comments.
@@ -42,11 +42,13 @@ RUN conda install -y pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoo | |||
# for graph | |||
RUN conda install -y pyg==2.0.1 -c pyg \ | |||
&& conda install -y rdkit=2021.09.4 -c conda-forge \ | |||
&& conda install -y nltk \ |
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.
nltk should be put at the back of the line (the NLP part below)?
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.
NLTK is used for generating features of some graph datasets.
local_update_steps: 1 | ||
total_round_num: 400 | ||
batch_or_epoch: 'epoch' | ||
client_num: 1 |
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.
why this baseline uses only 1 client?
DATA_LOAD_FUNCS = { | ||
'torchvision': load_torchvision_data, | ||
'torchtext': load_torchtext_data, | ||
'torchaudio': load_torchaudio_data, | ||
'torch_geometric': load_torch_geometric_data | ||
'torch_geometric': load_torch_geometric_data, | ||
'datasets': load_datasets_data, |
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.
the name "datasets" is too general, how about huggingface_datasets
federate: | ||
mode: standalone | ||
local_update_steps: 1 | ||
total_round_num: 400 |
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.
In centralized mode, the fine-tuning often takes only a few epochs. Maybe we can set the total_round_num to be in the order of dozens, e.g., 40?
Add more models and datasets from external packages.