Skip to content

fix megatron model & router replay #253

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
hjh0119 merged 2 commits into from
Jul 28, 2026
Merged

fix megatron model & router replay #253

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
d49f265
fix
hjh0119 Jul 28, 2026
466283e
Merge branch 'main' into fix-mg
hjh0119 Jul 28, 2026
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
7 changes: 4 additions & 3 deletions src/twinkle/model/megatron/megatron.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,7 +125,7 @@ def __init__(
self._model_path = HubOperation.download_model(model_id)
self.tokenizer_id = kwargs.get('tokenizer_id', self.model_id)
self._default_tokenizer = None
self.use_distributed_optimizer = kwargs.get('use_distributed_optimizer', True)
self.use_distributed_optimizer = kwargs.pop('use_distributed_optimizer', True)
self.variable_seq_lengths = kwargs.get('variable_seq_lengths', True)
torch_util.set_device()
self._try_init_process_group()
Expand Down Expand Up @@ -773,7 +773,7 @@ def _create_megatron_optimizer(self, **kwargs):
- weight_decay: Weight decay (default: 0.0)
- use_distributed_optimizer: Shard optimizer states (default: True)
- clip_grad: Gradient clipping threshold (default: 1.0)
- bf16: Use bf16 training (default: True)
- bf16 / fp16: precision flags (default: derived from self.mixed_precision)
- adam_beta1, adam_beta2, adam_eps: Adam parameters

Returns:
Expand All @@ -794,7 +794,8 @@ def _create_megatron_optimizer(self, **kwargs):
adam_beta2=kwargs.pop('adam_beta2', 0.999),
adam_eps=kwargs.pop('adam_eps', 1e-8),
clip_grad=kwargs.pop('clip_grad', 1.0),
bf16=kwargs.pop('bf16', True),
bf16=kwargs.pop('bf16', self.mixed_precision == 'bf16'),
fp16=kwargs.pop('fp16', self.mixed_precision == 'fp16'),
use_distributed_optimizer=self.use_distributed_optimizer,
overlap_param_gather=kwargs.pop('overlap_param_gather', False),
log_num_zeros_in_grad=kwargs.pop('log_num_zeros_in_grad', False),
Expand Down
8 changes: 3 additions & 5 deletions src/twinkle/processor/base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,8 +70,8 @@ def __init__(self,
self.framework = framework
self.process_pipeline = [
self.prepare_inputs,
self.align_routed_experts,
self.pad_cp,
self.align_routed_experts,
self.collate_fn,
self.to_transformers_dict,
self.add_extra_padding_free_args,
Expand Down Expand Up @@ -827,10 +827,8 @@ def align_routed_experts(self, inputs: Union[List[InputFeature], InputFeature],

def align_to(_input):
routed_experts = _input.get('routed_experts', None)
input_seq_len = _input.get('length', None)
if input_seq_len is None:
input_ids = _input.get('input_ids', None)
input_seq_len = input_ids.shape[1] if input_ids is not None else 0
input_ids = _input.get('input_ids', None)
input_seq_len = input_ids.shape[-1] if input_ids is not None else 0
if routed_experts is not None:
# The number of experts in the output can be 1 less than (prompt_length + response_token_count)
# This gap of 1 is expected
Expand Down
Loading