Skip to content

[T5] Bug correction & Refactor #8518

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.

Sign up for GitHub

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

Merged
patrickvonplaten merged 4 commits into from
Nov 13, 2020
Merged

[T5] Bug correction & Refactor #8518

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
802d75e
fix bug
patrickvonplaten Nov 12, 2020
148406d
T5 refactor
patrickvonplaten Nov 13, 2020
83676a2
refactor tf
patrickvonplaten Nov 13, 2020
3550934
apply sylvains suggestions
patrickvonplaten Nov 13, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 0 additions & 6 deletions src/transformers/configuration_t5.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,6 @@ class T5Config(PretrainedConfig):
def __init__(
self,
vocab_size=32128,
n_positions=512,

@patrickvonplaten patrickvonplaten Nov 13, 2020

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

T5 is not limited by a max number of positions since it uses relative pos embedding -> this param is never used in the Model File and should be deleted IMO: #8047

d_model=512,
d_kv=64,
d_ff=2048,
Expand All @@ -98,7 +97,6 @@ def __init__(
**kwargs,
)
self.vocab_size = vocab_size
self.n_positions = n_positions
self.d_model = d_model
self.d_kv = d_kv
self.d_ff = d_ff
Expand All @@ -112,10 +110,6 @@ def __init__(
self.layer_norm_epsilon = layer_norm_epsilon
self.initializer_factor = initializer_factor

@property
def max_position_embeddings(self):
return self.n_positions

@property
def hidden_size(self):
return self.d_model
Expand Down
4 changes: 1 addition & 3 deletions src/transformers/convert_t5_original_tf_checkpoint_to_pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,8 +17,6 @@

import argparse

import torch

from transformers import T5Config, T5Model, load_tf_weights_in_t5
from transformers.utils import logging

Expand All @@ -37,7 +35,7 @@ def convert_tf_checkpoint_to_pytorch(tf_checkpoint_path, config_file, pytorch_du

# Save pytorch-model
print("Save PyTorch model to {}".format(pytorch_dump_path))
torch.save(model.state_dict(), pytorch_dump_path)
model.save_pretrained(pytorch_dump_path)


if __name__ == "__main__":
Expand Down
Loading