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 11, 01:15 UTC
[ROCm] [Critical]: Remove unused variable (#31156)
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
TJian committed
Dec 23, 2025 at 01:28 UTC
022f3cea5327cc720a325c50931e1edcfdf2d32b
1 file changed
-1
csrc/fused_qknorm_rope_kernel.cu
-1
@@ -405,7 +405,6 @@ void fused_qk_norm_rope(
405
qkv.scalar_type() == k_weight.scalar_type(),
406
"qkv, q_weight and k_weight must have the same dtype");
407
408
- int64_t rotary_dim = cos_sin_cache.size(1);
408
int64_t num_tokens = qkv.size(0);
409
TORCH_CHECK(position_ids.size(0) == num_tokens,
410
"Number of tokens in position_ids must match QKV");