Skip to content

Commit 7f1a038

Browse files
author
bzgoogle
committed
correct # layer
1 parent d7db532 commit 7f1a038

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

tpu_commons/models/jax/deepseek_v3.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -45,7 +45,7 @@ def __init__(self,
4545
self.vllm_config = vllm_config
4646
self.rng = nnx.Rngs(rng)
4747

48-
num_layers: int = 5
48+
num_layers: int = 61
4949
num_local_experts: int = 256
5050

5151
vocab_size: int = 129280

0 commit comments

Comments
 (0)