Skip to content

[deepseek] fix mismatch of model embeddings vs tokenizer size (128815), sync to 671B vocab size, avoids cudaErrorAssert and run termination. #1495

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

Closed
lessw2020 wants to merge 2 commits into from
Closed

[deepseek] fix mismatch of model embeddings vs tokenizer size (128815), sync to 671B vocab size, avoids cudaErrorAssert and run termination. #1495

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
2564f43
fix mismatch of model embeddings vs tokenizer size (128815)
lessw2020 Jul 30, 2025
5d36a43
move to 671B vocab sizer for consistency (129280)
lessw2020 Jul 30, 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
4 changes: 2 additions & 2 deletions torchtitan/models/deepseek_v3/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,7 @@
attn_mask_type="block_causal",
),
"16B": DeepSeekV3ModelArgs(
vocab_size=102400,
vocab_size=129280,
dim=2048,
inter_dim=10944,
moe_inter_dim=1408,
Expand All @@ -88,7 +88,7 @@
mscale=0.70,
),
"236B": DeepSeekV3ModelArgs(
vocab_size=102400,
vocab_size=129280,
dim=5120,
inter_dim=12288,
moe_inter_dim=1536,
Expand Down
2 changes: 1 addition & 1 deletion torchtitan/models/deepseek_v3/model/args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@ class DeepSeekV3ModelArgs(BaseModelArgs):
max_batch_size: int = 8
max_seq_len: int = 4096 * 4
dtype: Literal["bf16", "fp8"] = "bf16"
vocab_size: int = 102400
vocab_size: int = 129280
dim: int = 2048
inter_dim: int = 10944
moe_inter_dim: int = 1408
Expand Down
Loading