We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent b33f11b commit 6ecb2bcCopy full SHA for 6ecb2bc
vllm/model_executor/models/llama4.py
@@ -375,7 +375,8 @@ def is_fp4_weight(weight):
375
new_loaded_weight = new_loaded_weight[shard_idx]
376
377
# Only transpose for non-FP4 weights
378
- # FP4 weights are already in the correct format and shouldn't be transposed here.
+ # FP4 weights are already in the correct format and
379
+ # shouldn't be transposed here.
380
if not is_fp4_weight(new_loaded_weight):
381
new_loaded_weight = new_loaded_weight.transpose(-1, -2)
382
0 commit comments