refactor: unify runner execution and weight residency (#2321)

This commit is contained in:
leejet authored and GitHub committed 2026-11-10 23:45:01 +08:00
1 parent 56070ea034
commit ddc367ab5d
15 files changed
+1679 -1253

No files matched your search

+1 -1
View File
@@ -576,7 +576,7 @@ namespace Qwen {
result = run(cache);
}
if (result.empty() && last_compute_status() == GGML_STATUS_ALLOC_FAILED &&
(cache.mode != QwenImage21PrefixCache::Mode::NONE || !cache_.empty())) {
(cache.mode != QwenImage21PrefixCache::Mode::NONE || has_cache_tensors())) {
// The failed graph has ended before persistent inputs are released.
free_cache_ctx_and_buffer();
prefix_cache_disabled = true;