Skip to content

[Tests][e2e] Release memory before running vLLM #2375

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kylesayrs merged 3 commits into from
Feb 19, 2026
Merged

[Tests][e2e] Release memory before running vLLM #2375

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
695a43f
release memory after the test
dsikka Feb 17, 2026
f215109
Merge branch 'main' into e2e_cleanup
dsikka Feb 17, 2026
fabdfc5
Merge branch 'main' into e2e_cleanup
kylesayrs Feb 19, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 1 addition & 2 deletions tests/e2e/vLLM/configs/w4a16_grouped_quant_asym_awq.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,5 +5,4 @@ recipe: tests/e2e/vLLM/recipes/WNA16/recipe_w4a16_awq_asym.yaml
dataset_id: HuggingFaceH4/ultrachat_200k
dataset_split: train_sft
scheme: W4A16_weight_asym_awq
save_dir: TinyLlama-1.1B-Chat-v1.0-w4a16-asym-awq
gpu_memory_utilization: 0.85
save_dir: TinyLlama-1.1B-Chat-v1.0-w4a16-asym-awq
1 change: 0 additions & 1 deletion tests/e2e/vLLM/configs/w4a16_grouped_quant_sym_awq.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,4 +6,3 @@ dataset_id: HuggingFaceH4/ultrachat_200k
dataset_split: train_sft
scheme: W4A16_weight_sym_awq
save_dir: TinyLlama-1.1B-Chat-v1.0-w4a16-sym-awq
gpu_memory_utilization: 0.85
13 changes: 13 additions & 0 deletions tests/e2e/vLLM/test_vllm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,11 +1,14 @@
import gc
import os
import re
import shutil
import sys
import time
from pathlib import Path

import pandas as pd
import pytest
import torch
import yaml
from huggingface_hub import HfApi
from loguru import logger
Expand Down Expand Up @@ -138,6 +141,16 @@ def save_compressed_model(self):
fp.write(recipe_yaml_str)
session.reset()

# Release GPU memory before running vLLM
del self.oneshot_model
del self.tokenizer

gc.collect()
torch.cuda.empty_cache()
torch.cuda.synchronize()
# Give GPU time to fully release memory
time.sleep(2)

if SKIP_HF_UPLOAD.lower() != "yes":
logger.info("================= UPLOADING TO HUB ======================")

Expand Down