Skip to content

fix: ensure that we check for pad_token and not assume pad_token==eos… #120

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
parthchadha merged 2 commits into from
Apr 2, 2025
Merged

fix: ensure that we check for pad_token and not assume pad_token==eos… #120

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5669216
fix: ensure that we check for pad_token and not assume pad_token==eos…
parthchadha Apr 2, 2025
2983ca7
Add an assertion for checking if tokenizer has pad_token_id assigned
parthchadha Apr 2, 2025
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: 3 additions & 3 deletions nemo_reinforcer/algorithms/grpo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -472,7 +472,7 @@ def grpo_train(
# Convert LLMMessageLogType to FlatMessagesType for generation
batched_flat, input_lengths = batched_message_log_to_flat_message(
repeated_batch["message_log"],
pad_value_dict={"token_ids": tokenizer.eos_token_id},
pad_value_dict={"token_ids": tokenizer.pad_token_id},
Comment thread
terrykong marked this conversation as resolved.
)
input_ids = batched_flat["token_ids"]
# Create generation-specific input structure
Expand Down Expand Up @@ -547,7 +547,7 @@ def grpo_train(
# Convert updated LLMMessageLogType to FlatMessagesType for training
flat_messages, input_lengths = batched_message_log_to_flat_message(
repeated_batch["message_log"],
pad_value_dict={"token_ids": tokenizer.eos_token_id},
pad_value_dict={"token_ids": tokenizer.pad_token_id},
)

# Create training data from flattened messages
Expand Down Expand Up @@ -704,7 +704,7 @@ def validate(
# Convert LLMMessageLogType to FlatMessagesType for generation
batched_flat, input_lengths = batched_message_log_to_flat_message(
val_batch["message_log"],
pad_value_dict={"token_ids": tokenizer.eos_token_id},
pad_value_dict={"token_ids": tokenizer.pad_token_id},
)
# Extract input IDs
input_ids = batched_flat["token_ids"]
Expand Down
4 changes: 2 additions & 2 deletions nemo_reinforcer/algorithms/sft.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,7 +243,7 @@ def validate(

cat_and_padded, input_lengths = batched_message_log_to_flat_message(
val_batch["message_log"],
pad_value_dict={"token_ids": tokenizer.eos_token_id},
pad_value_dict={"token_ids": tokenizer.pad_token_id},
)

val_data: BatchedDataDict = BatchedDataDict(
Expand Down Expand Up @@ -356,7 +356,7 @@ def sft_train(

cat_and_padded, input_lengths = batched_message_log_to_flat_message(
batch["message_log"],
pad_value_dict={"token_ids": tokenizer.eos_token_id},
pad_value_dict={"token_ids": tokenizer.pad_token_id},
)

train_data: BatchedDataDict = BatchedDataDict(
Expand Down
5 changes: 5 additions & 0 deletions nemo_reinforcer/models/generation/interfaces.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,11 @@ def verify_right_padding(
f"data must be a BatchedDataDict, got type: {type(data)}"
)

assert pad_value is not None, (
"Tokenizer does not have a pad token assigned. \n"
"If the default tokenizer does not have a pad token, you can assign it the value of eos token by tokenizer.pad_token = tokenizer.eos_token"
)

# Determine which type of data we're dealing with
if "input_ids" in data and "input_lengths" in data:
# GenerationDatumSpec
Expand Down