Skip to content

GitLab

Unverified Commit 3e34e900 authored by narutolhy's avatar narutolhy Committed by GitHub
Browse files

Fix: sync prepare_fp8_layer_for_marlin with latest vllm changes (#7648)

parent 7349717e
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
python/sglang/srt/layers/quantization/compressed_tensors/schemes/compressed_tensors_w8a16_fp8.py
View file @ 3e34e900
...@@ -76,7 +76,7 @@ class CompressedTensorsW8A16Fp8(CompressedTensorsScheme): ...@@ -76,7 +76,7 @@ class CompressedTensorsW8A16Fp8(CompressedTensorsScheme):
layer.input_scale = torch.nn.Parameter( layer.input_scale = torch.nn.Parameter(
layer.input_scale.data, requires_grad=False layer.input_scale.data, requires_grad=False
) )
prepare_fp8_layer_for_marlin(layer, strategy="channel") prepare_fp8_layer_for_marlin(layer, size_k_first=True)
def create_weights( def create_weights(
self, self,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment