A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 23:30 UTC
[Bugfix] Fix invalid rotary embedding unit test (#13431)
Signed-off-by: Liangfu Chen <liangfc@amazon.com>
Liangfu Chen committed
Feb 18, 2025 at 03:52 UTC
38094584566b89210a6f72a408eba1fae43c3d81
1 file changed
+1
-1
tests/kernels/test_rotary_embedding.py
+1
-1
@@ -41,7 +41,7 @@ def test_rotary_embedding_opcheck(dist_init, device, max_position,
41
is_neox_style, rotary_dim, head_size,
42
seq_len):
43
batch_size = 1
44
- base = 0
44
+ base = 10000
45
num_heads = 7
46
rot = RotaryEmbedding(head_size, rotary_dim, max_position, base,
47
is_neox_style, torch.float32)