From 7387fae2ba5252291ece8fef29990383df4b1c7f Mon Sep 17 00:00:00 2001 From: danielhanchen Date: Mon, 31 Aug 2026 10:06:10 +0000 Subject: [PATCH 1/4] unsloth: pin ggml-org#27941, #152 and #154 ggml-org#27941 at 02eb201d, no longer a draft and MERGEABLE upstream. It fixes four qwen4exp correctness defects and is the only account of the Flash-Next reporter who is on Vulkan only, where the allocator variable is never set. The previous pin PR #148 carried the stale 8161d117 and is closed. #152 at 258345ef and #154 at 31e432e7 both merge onto b10708 on their own. #137 is already pinned at 4e1865e3, which is still its head, so no change. #142 and #144 are NOT pinned here: both predate an upstream refactor that folded lazy_mode and model_shared into a lazy struct, so they conflict on src/llama.cpp and src/llama-model-loader.cpp against b10708. b10708 has 'ml.lazy.mode = params.lazy_mode' where those branches still write 'ml.lazy_mode' and 'ml.model_shared'. They need rebasing onto current upstream before they can be pinned; pinning them now would fail the resolve. --- scripts/unsloth/pr-set.json | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/scripts/unsloth/pr-set.json b/scripts/unsloth/pr-set.json index 3e69aa63d3de..9e5c9703a082 100644 --- a/scripts/unsloth/pr-set.json +++ b/scripts/unsloth/pr-set.json @@ -27,6 +27,9 @@ "https://github.com/unslothai/llama.cpp/pull/137/commits/4e1865e34ec5f6ca39403215c89129c13731be70", "https://github.com/unslothai/llama.cpp/pull/158/commits/abfc45b9cb21eae4848cb82196e659f42c9a8341", "https://github.com/unslothai/llama.cpp/pull/157/commits/6c6da89266ba7839d825c9997782af4f4d26b81b", - "https://github.com/unslothai/llama.cpp/pull/149/commits/b65a2dce12c14a489e19a059cb6ee59112f1b733" + "https://github.com/unslothai/llama.cpp/pull/149/commits/b65a2dce12c14a489e19a059cb6ee59112f1b733", + "https://github.com/ggml-org/llama.cpp/pull/27941/commits/02eb201def6cecb12c25b7b4613d44e546acabda", + "https://github.com/unslothai/llama.cpp/pull/152/commits/258345efa640eb099eb1af3c9ee8f8e6e8e7b0d3", + "https://github.com/unslothai/llama.cpp/pull/154/commits/31e432e758f8cc4b2c5f27902721500173bf39db" ] } From 4991f8637f3d6bcdfa1b3a7d031bb6b199345d33 Mon Sep 17 00:00:00 2001 From: danielhanchen Date: Mon, 31 Aug 2026 10:53:41 +0000 Subject: [PATCH 2/4] unsloth: repin ggml-org#27941 to 6b2b85cf The PR moved on at 10:44Z, after 02eb201d was pinned. Verified to merge onto the current base tag: 6 files changed, 480 insertions, 138 deletions. --- scripts/unsloth/pr-set.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/unsloth/pr-set.json b/scripts/unsloth/pr-set.json index 9e5c9703a082..b0388bbd877c 100644 --- a/scripts/unsloth/pr-set.json +++ b/scripts/unsloth/pr-set.json @@ -28,7 +28,7 @@ "https://github.com/unslothai/llama.cpp/pull/158/commits/abfc45b9cb21eae4848cb82196e659f42c9a8341", "https://github.com/unslothai/llama.cpp/pull/157/commits/6c6da89266ba7839d825c9997782af4f4d26b81b", "https://github.com/unslothai/llama.cpp/pull/149/commits/b65a2dce12c14a489e19a059cb6ee59112f1b733", - "https://github.com/ggml-org/llama.cpp/pull/27941/commits/02eb201def6cecb12c25b7b4613d44e546acabda", + "https://github.com/ggml-org/llama.cpp/pull/27941/commits/6b2b85cfbf8a5c612e7d3d4c0ef222e1e58de1c5", "https://github.com/unslothai/llama.cpp/pull/152/commits/258345efa640eb099eb1af3c9ee8f8e6e8e7b0d3", "https://github.com/unslothai/llama.cpp/pull/154/commits/31e432e758f8cc4b2c5f27902721500173bf39db" ] From 3028768f31bd561092248bd69f3157945548e33d Mon Sep 17 00:00:00 2001 From: danielhanchen Date: Mon, 31 Aug 2026 11:00:29 +0000 Subject: [PATCH 3/4] unsloth: pin #144, rebased, and leave #142 out as contained in it #144 rebased onto b10709 at 6fc8df13. Two of its nine commits were dropped rather than carried: 'qwen4exp: fix QSA correctness defects and harden metadata loading' and the comment tidy on top of it. That work is what ggml-org#27941 supersedes, and keeping both copies is what made the two pins conflict in llama-kv-cells.h and llama-memory-hybrid-idx.cpp. Listed after #27941 so the upstream version of that work lands first. #142 is deliberately NOT pinned: its single commit is byte-identical to #144's c7bd6f207 apart from the lazy API spelling, so #144 already contains it and pinning both would apply the same change twice. Verified: b10709 + #27941 + #144 merges clean, 24 files, +612/-94. --- scripts/unsloth/pr-set.json | 1 + 1 file changed, 1 insertion(+) diff --git a/scripts/unsloth/pr-set.json b/scripts/unsloth/pr-set.json index b0388bbd877c..87c097d86597 100644 --- a/scripts/unsloth/pr-set.json +++ b/scripts/unsloth/pr-set.json @@ -29,6 +29,7 @@ "https://github.com/unslothai/llama.cpp/pull/157/commits/6c6da89266ba7839d825c9997782af4f4d26b81b", "https://github.com/unslothai/llama.cpp/pull/149/commits/b65a2dce12c14a489e19a059cb6ee59112f1b733", "https://github.com/ggml-org/llama.cpp/pull/27941/commits/6b2b85cfbf8a5c612e7d3d4c0ef222e1e58de1c5", + "https://github.com/unslothai/llama.cpp/pull/144/commits/6fc8df13a1677759a06ae229155d697c5d725ea7", "https://github.com/unslothai/llama.cpp/pull/152/commits/258345efa640eb099eb1af3c9ee8f8e6e8e7b0d3", "https://github.com/unslothai/llama.cpp/pull/154/commits/31e432e758f8cc4b2c5f27902721500173bf39db" ] From a76f5561eaf2fa16defb77491a5ea9b1c6d34cd3 Mon Sep 17 00:00:00 2001 From: danielhanchen Date: Mon, 31 Aug 2026 11:05:05 +0000 Subject: [PATCH 4/4] unsloth: repin ggml-org#27754 to 949f7efb The PR moved on at 10:39Z; 5796547f was pinned by #159 earlier today. Verified to merge onto b10709: 44 files changed, 2673 insertions, 38 deletions. Every entry in the set is now at its PR's current head. --- scripts/unsloth/pr-set.json | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/unsloth/pr-set.json b/scripts/unsloth/pr-set.json index 87c097d86597..96215c47cb84 100644 --- a/scripts/unsloth/pr-set.json +++ b/scripts/unsloth/pr-set.json @@ -23,7 +23,7 @@ "https://github.com/unslothai/llama.cpp/pull/70/commits/edfd4c1a3b7a653303a85257ddac2a1f3ce39a2f", "https://github.com/unslothai/llama.cpp/pull/91/commits/c86ed269986f2dced6325c5c58bda966a2e2ead1", "https://github.com/unslothai/llama.cpp/pull/95/commits/3db8cb5b2e9bf291057b9f19960e8601a162da81", - "https://github.com/ggml-org/llama.cpp/pull/27754/commits/5796547f37f5943513dfa130065ec88f9e30e0f5", + "https://github.com/ggml-org/llama.cpp/pull/27754/commits/949f7efb097eb20ef36fecdb1afaebff9a4ae7ed", "https://github.com/unslothai/llama.cpp/pull/137/commits/4e1865e34ec5f6ca39403215c89129c13731be70", "https://github.com/unslothai/llama.cpp/pull/158/commits/abfc45b9cb21eae4848cb82196e659f42c9a8341", "https://github.com/unslothai/llama.cpp/pull/157/commits/6c6da89266ba7839d825c9997782af4f4d26b81b",