Skip to content

Commit 8cfb5ec

Browse files
dsikkadsikka
authored
[LM Eval] Move vl configs to a different folder (#2122)
SUMMARY: This will make it easier to run the lm-eval tests in two different jobs, one focused on VL models and the other focused on a normal decoder model
1 parent ceed4df commit 8cfb5ec

File tree

4 files changed

+1
-1
lines changed

4 files changed

+1
-1
lines changed

‎tests/lmeval/configs/w4a16_awq_sym.yaml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
cadence: "weekly"
22
model: meta-llama/Meta-Llama-3-8B-Instruct
3-
scheme: W4A16
3+
scheme: W4A16_awq_sym
44
recipe: tests/e2e/vLLM/recipes/WNA16/recipe_w4a16_awq_sym.yaml
55
dataset_id: HuggingFaceH4/ultrachat_200k
66
dataset_split: train_sft

tests/lmeval/configs/vl_fp8_dynamic_per_token.yaml renamed to tests/lmeval/vl_configs/vl_fp8_dynamic_per_token.yaml

File renamed without changes.

tests/lmeval/configs/vl_int8_w8a8_dynamic_per_token.yaml renamed to tests/lmeval/vl_configs/vl_int8_w8a8_dynamic_per_token.yaml

File renamed without changes.

tests/lmeval/configs/vl_w4a16_actorder_weight.yaml renamed to tests/lmeval/vl_configs/vl_w4a16_actorder_weight.yaml

File renamed without changes.

0 commit comments

Comments
 (0)