[examples] add --enable-cpu-offload args (#930)

Signed-off-by: David Chen <530634352@qq.com>
This commit is contained in:
WeiQing Chen authored and GitHub committed 2026-01-23 22:43:00 +08:00
1 parent 28a6597c4c
commit 3d9cd16df7
8 files changed
+28

No files matched your search

@@ -61,6 +61,7 @@ Key arguments:
- `--num_inference_steps`: diffusion sampling steps (more steps = higher quality, slower).
- `--height/--width`: output resolution (defaults 1024x1024).
- `--output`: path to save the generated PNG.
- `--enable-cpu-offload`: enable CPU offloading for diffusion models.
> ℹ️ Qwen-Image currently publishes best-effort presets at `1328x1328`, `1664x928`, `928x1664`, `1472x1140`, `1140x1472`, `1584x1056`, and `1056x1584`. Adjust `--height/--width` accordingly for the most reliable outcomes.
@@ -96,6 +96,11 @@ def parse_args() -> argparse.Namespace:
action="store_true",
help="Disable torch.compile and force eager execution.",
)
parser.add_argument(
"--enable-cpu-offload",
action="store_true",
help="Enable CPU offloading for diffusion models.",
)
parser.add_argument(
"--tensor_parallel_size",
type=int,
@@ -162,6 +167,7 @@ def main():
cache_config=cache_config,
parallel_config=parallel_config,
enforce_eager=args.enforce_eager,
enable_cpu_offload=args.enable_cpu_offload,
)
if profiler_enabled: