Skip to content

feat(train): freeze target components and support param-list in BF16Optimizer #561

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

Open
curnane-lab wants to merge 1 commit into
base: main
Choose a base branch
from
+25 −3
Open

feat(train): freeze target components and support param-list in BF16Optimizer #561

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
16 changes: 15 additions & 1 deletion scripts/train_dflash.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -438,6 +438,13 @@ def main():
loss_decay_gamma=args.loss_decay_gamma,
)

# Freeze target components: only train draft model
for param in target_components.lm_head.parameters():
param.requires_grad = False
for param in target_components.embed_tokens.parameters():
param.requires_grad = False
print_on_rank0("Frozen target_lm_head and target_embed_tokens")

dflash_model = FSDP(
dflash_model,
use_orig_params=True,
Expand All @@ -452,8 +459,15 @@ def main():
start_epoch = ckpt_info[0]
global_step = ckpt_info[1]

# BF16Optimizer must operate on the FSDP-wrapped model's draft parameters
# to ensure gradient synchronization works correctly with use_orig_params=True
draft_params = []
for name, param in dflash_model.named_parameters():
if param.requires_grad and "draft_model." in name:
draft_params.append(param)

optimizer = BF16Optimizer(
draft_model,
draft_params,
lr=args.learning_rate,
max_grad_norm=args.max_grad_norm,
warmup_ratio=args.warmup_ratio,
Expand Down
12 changes: 10 additions & 2 deletions specforge/optimizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,8 +18,16 @@ def __init__(
# TODO: We should make these parameters configurable
# These magic numbers: weight_decay=0.0, max_grad_norm=0.5, total_steps=800k, warmup_steps=12k are copied from
# https://github.com/SafeAILab/EAGLE/blob/main/eagle/traineagle3/ds_config.json
self.model = model
self.model_params = [p for p in model.parameters() if p.requires_grad]
if isinstance(model, list):
self.model = None
self.model_params = [
p for p in model if p.requires_grad
]
else:
self.model = model
self.model_params = [
p for p in model.parameters() if p.requires_grad
]
self.max_grad_norm = max_grad_norm
self.fp32_params = [
p.detach().clone().to(torch.float32) for p in self.model_params
Expand Down