Skip to content

Commit 8e7484f

Browse files
stick to regular 2:4 sparsity, avoid marlin kernel
Signed-off-by: Brian Dellabetta <[email protected]>
1 parent 35f1d50 commit 8e7484f

File tree

3 files changed

+8
-18
lines changed

3 files changed

+8
-18
lines changed
Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
cadence: "nightly"
2+
test_type: "regression"
3+
model: Qwen/Qwen2.5-0.5B
4+
recipe: tests/e2e/vLLM/recipes/Sparse_2of4/recipe_sparse_2of4.yaml
5+
scheme: sparse2of4_only
6+
dataset_id: garage-bAInd/Open-Platypus
7+
dataset_split: train
8+
save_compressed: False

tests/e2e/vLLM/configs/w4a16_2of4_channel_quant_qwen.yaml

Lines changed: 0 additions & 9 deletions
This file was deleted.

tests/e2e/vLLM/configs/w4a16_2of4_grouped_quant_qwen.yaml

Lines changed: 0 additions & 9 deletions
This file was deleted.

0 commit comments

Comments
 (0)