From d40e5f6217f01eeb27dbfa1b0b2aa22740181224 Mon Sep 17 00:00:00 2001 From: danielhanchen Date: Wed, 16 Sep 2026 13:49:18 +0000 Subject: [PATCH] Repin GLM-5-Next onto the head carrying the indexer softmax fix unslothai#214 landed on glm5next/upstream after #216 was cut, so e2738e07 is no longer the head. 86ebfef2 is that squash on top of it: the k-pool gate logits are reshaped to 2D before ggml_soft_max so n_new_max stops mapping to gridDim.y, which CUDA caps at 65535 and which aborted the launch at n_kv >= 262144 with kpool = 4. Replayed the resolve loop on b10994 with the new pin: 11 clean, 2 additive, 0 hard fails; merge_checks clean, all 13 pins intact, llama + mtmd compile gate passed, test-llama-archs 316 rows 0 failures. --- 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 2502348f0e16..0de8aa2db2a1 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/883f2c9ba78f3847148454adf025da29385fff3e", "https://github.com/unslothai/llama.cpp/pull/61/commits/46cbf0e95786fe8f5b7c0e86d57aaf8f8eceea7f", "https://github.com/unslothai/llama.cpp/pull/95/commits/3db8cb5b2e9bf291057b9f19960e8601a162da81", - "https://github.com/ggml-org/llama.cpp/pull/27754/commits/e2738e070b0f102a461cdd7d5a1706e9be06d848", + "https://github.com/ggml-org/llama.cpp/pull/27754/commits/86ebfef2c6a0f3359a2a07d2c215d61b0fa885c9", "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",