fix(fish-speech): use CUDA toolkit ptxas

Prefer an explicitly configured Triton assembler, otherwise use the executable ptxas from CUDA_HOME so torch.compile can target GPU architectures newer than Triton bundled tooling.

Assisted-by: Codex:gpt-5

.docker/llama-cpp-build-target.sh

@@ -8,8 +8,15 @@ build_type=${2-}
 # ggml-cpu/arch/x86/repack.cpp at -march=sapphirerapids: the job sits on that one
 # translation unit until GitHub kills it at 6h. gcc builds the same file in
 # seconds, so only the SYCL images have to give up the CPU variant matrix.
+#
+# ROCm runs out of the same 6h budget for a different reason: volume, not a
+# stall. hipcc compiles ggml's HIP kernels once per entry in AMDGPU_TARGETS,
+# which is eleven architectures (gfx908 through gfx1201), and the CPU variant
+# matrix lands on top of that. The job built in 2h27m before it was added and
+# has been killed at exactly 6h00m on every run since, so no ROCm llama-cpp
+# image has been published since 2026-08-01.
 case "$build_type" in
-  sycl*)
+  sycl*|hipblas*)
     echo llama-cpp-fallback
     exit 0
     ;;

backend/cpp/ds4/CMakeLists.txt

@@ -69,7 +69,15 @@ target_include_directories(hw_grpc_proto PUBLIC ${CMAKE_CURRENT_BINARY_DIR})
 
 set(DS4_OBJS "${DS4_DIR}/ds4.o")
 if(DS4_GPU STREQUAL "cuda")
-    list(APPEND DS4_OBJS "${DS4_DIR}/ds4_cuda.o")
+    list(APPEND DS4_OBJS
+        "${DS4_DIR}/ds4_cuda.o"
+        "${DS4_DIR}/cuda/mmq/ds4_ggml_stubs.o"
+        "${DS4_DIR}/cuda/mmq/ds4_mmq.o"
+        "${DS4_DIR}/cuda/mmq/ds4_mmq_d2r.o"
+        "${DS4_DIR}/cuda/mmq/quantize.o"
+        "${DS4_DIR}/cuda/mmq/mmid.o"
+        "${DS4_DIR}/cuda/mmq/mmvq.o"
+        "${DS4_DIR}/cuda/mmq/ds4_repack.o")
 elseif(DS4_GPU STREQUAL "metal")
     list(APPEND DS4_OBJS "${DS4_DIR}/ds4_metal.o")
 elseif(DS4_GPU STREQUAL "cpu")

backend/cpp/ds4/Makefile

@@ -1,10 +1,10 @@
 # ds4 backend Makefile.
 #
-# Upstream pin lives below as DS4_VERSION?=54b36ed9ba42da31b24f2d1a5feb075c2475dbb1
+# Upstream pin lives below as DS4_VERSION?=b7e9f0091139999b6c070a57590c447c5741da5c
 # (.github/bump_deps.sh) can find and update it - matches the
 # llama-cpp / ik-llama-cpp / turboquant convention.
 
-DS4_VERSION?=54b36ed9ba42da31b24f2d1a5feb075c2475dbb1
+DS4_VERSION?=b7e9f0091139999b6c070a57590c447c5741da5c
 DS4_REPO?=https://github.com/antirez/ds4
 
 CURRENT_MAKEFILE_DIR := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
@@ -23,7 +23,9 @@ CMAKE_ARGS ?= -DCMAKE_BUILD_TYPE=Release
 # are shared by every GPU mode, so append them unconditionally below.
 ifeq ($(BUILD_TYPE),cublas)
     CMAKE_ARGS += -DDS4_GPU=cuda
-    DS4_OBJ_TARGET := ds4.o ds4_cuda.o ds4_distributed.o ds4_tp.o ds4_ssd.o ds4_layer_pack.o
+    DS4_OBJ_TARGET := ds4.o ds4_cuda.o ds4_distributed.o ds4_tp.o ds4_ssd.o ds4_layer_pack.o \
+        cuda/mmq/ds4_ggml_stubs.o cuda/mmq/ds4_mmq.o cuda/mmq/ds4_mmq_d2r.o \
+        cuda/mmq/quantize.o cuda/mmq/mmid.o cuda/mmq/mmvq.o cuda/mmq/ds4_repack.o
 else ifeq ($(UNAME_S),Darwin)
     CMAKE_ARGS += -DDS4_GPU=metal
     DS4_OBJ_TARGET := ds4.o ds4_metal.o ds4_distributed.o ds4_tp.o ds4_ssd.o ds4_layer_pack.o
@@ -55,7 +57,7 @@ ds4:
 # the right per-platform compile flags (Objective-C/Metal on Darwin, nvcc on Linux+CUDA).
 ds4/ds4.o: ds4
 ifeq ($(BUILD_TYPE),cublas)
-	+$(MAKE) -C ds4 ds4.o ds4_cuda.o ds4_distributed.o ds4_tp.o ds4_ssd.o ds4_layer_pack.o
+	+$(MAKE) -C ds4 $(DS4_OBJ_TARGET)
 else ifeq ($(UNAME_S),Darwin)
 	+$(MAKE) -C ds4 ds4.o ds4_metal.o ds4_distributed.o ds4_tp.o ds4_ssd.o ds4_layer_pack.o
 else

backend/go/vllm-cpp/Makefile

@@ -11,7 +11,30 @@ JOBS?=$(shell nproc --ignore=1 2>/dev/null || sysctl -n hw.ncpu 2>/dev/null || e
 
 # vllm.cpp version
 VLLM_CPP_REPO?=https://github.com/mudler/vllm.cpp
-VLLM_CPP_VERSION?=a42b8187caff02c570c28e19e4dc2b1d7f55ed14
+VLLM_CPP_VERSION?=9d1fad3cde0acb95eb0bb0a1025f40a0eb614147
+
+# MLX GEMM provider (darwin/metal only; see the metal branch below for why).
+# Consumed as the prebuilt pip wheel: building MLX from source needs `xcrun
+# metal`, i.e. a full Xcode the macOS runners do not have, while the wheel ships
+# include/, lib/libmlx.dylib and the compiled mlx.metallib ready to link.
+#
+# DEFAULT ON, but ONLY because VLLM_CPP_VERSION above is pinned at or past
+# vllm.cpp 89c46aeb, which SHAPE-GATES the provider to prefill. The ordering is
+# load-bearing, not incidental:
+#
+#   pin >= 89c46aeb, MLX on   -> 99.1% of MLX-LM  (gated: prefill only)
+#   pin <  89c46aeb, MLX on   -> ~51%             (ungated: it also takes decode)
+#
+# MLX's steel GEMM wins prefill (537 ms TTFT against 602) and loses decode badly,
+# because the provider pays an mx::eval sync plus an output memcpy per call and
+# decode makes ~112 calls per TOKEN. Ungated it does both; gated it does only the
+# good half. So if this pin is ever moved BACKWARDS, this default must go with it.
+VLLM_CPP_MLX?=on
+MLX_VERSION?=0.29.4
+MLX_VENV?=$(abspath ./mlx-venv)
+# Resolved lazily (recursive `=`, not `:=`): the glob only matches once the venv
+# target has run, and the interpreter version in the path varies per runner.
+MLX_ROOT=$(shell echo $(MLX_VENV)/lib/python*/site-packages/mlx)
 
 # The backend consumes only the stable C ABI (libvllm + include/vllm.h), so the
 # server, examples and tests of the engine are never built here.
@@ -49,6 +72,23 @@ else ifeq ($(BUILD_TYPE),vulkan)
 	CMAKE_ARGS+=-DVLLM_CPP_VULKAN=ON -DVLLM_CPP_CUDA=OFF
 else ifeq ($(BUILD_TYPE),metal)
 	CMAKE_ARGS+=-DVLLM_CPP_METAL=ON
+	# The optional MLX GEMM provider. vllm.cpp keeps it OFF by default because it
+	# is a ~19 MB libmlx.dylib plus a ~105 MB mlx.metallib, and upstream's
+	# position is that it must earn that cost by measurement. It does, on the
+	# only hardware this build targets: measured on an Apple M4 against the
+	# native MSL GEMM in the SAME binary (arms toggled by
+	# VT_OP_PROVIDER_DISABLE=mlx), Qwen3-1.7B-bf16 p=512 g=128, it is 1.5x to
+	# 2.2x aggregate throughput and 2x to 3x faster TTFT, at equal peak memory
+	# and bit-identical output on every parity shape. See vllm.cpp
+	# docs/BENCHMARKS.md "MLX GEMM provider A/B on Apple M4".
+	#
+	# MLX delegates the dense GEMM ONLY: kPagedAttention stays vllm.cpp's own
+	# kernel, because MLX has no paged-KV primitive at all.
+	#
+	# Set VLLM_CPP_MLX=off for a Metal build without it (smaller image, slower).
+	ifeq ($(VLLM_CPP_MLX),on)
+		MLX_ENABLED=1
+	endif
 else
 	CMAKE_ARGS+=-DVLLM_CPP_CUDA=OFF
 endif
@@ -68,10 +108,35 @@ sources/vllm.cpp:
 	git fetch --depth 1 origin $(VLLM_CPP_VERSION) && \
 	git checkout FETCH_HEAD
 
-$(LIB): sources/vllm.cpp
+ifeq ($(MLX_ENABLED),1)
+# A stamp FILE, not a phony target: a phony prerequisite is always "newer" than
+# $(LIB) and would re-link libvllm on every invocation. Keyed on the version so
+# a MLX_VERSION bump reinstalls instead of silently reusing the old wheel.
+MLX_STAMP=$(MLX_VENV)/.mlx-$(MLX_VERSION).stamp
+MLX_CMAKE_ARGS=-DVLLM_CPP_MLX=ON -DMLX_ROOT=$(MLX_ROOT)
+
+$(MLX_STAMP):
+	@if [ ! -x "$(MLX_VENV)/bin/pip" ]; then \
+		python3 -m venv "$(MLX_VENV)" || { echo "vllm-cpp: python3 with venv is required to build the MLX provider; pass VLLM_CPP_MLX=off to build Metal without it" >&2; exit 1; }; \
+	fi
+	"$(MLX_VENV)"/bin/pip install --quiet --disable-pip-version-check "mlx==$(MLX_VERSION)"
+	@# Resolved in the SHELL, not by $(MLX_ROOT): make expands a whole recipe
+	@# before running its first line, so the glob would still be unmatched here.
+	@# Every later use (the cmake args, package.sh) expands after this target has
+	@# completed, where $(MLX_ROOT) does resolve.
+	@root=$$(echo "$(MLX_VENV)"/lib/python*/site-packages/mlx); \
+	test -f "$$root/lib/libmlx.dylib" -a -f "$$root/include/mlx/array.h" || \
+		{ echo "vllm-cpp: mlx==$(MLX_VERSION) did not provide lib/libmlx.dylib + include/mlx/array.h under $$root" >&2; exit 1; }
+	touch $@
+else
+MLX_STAMP=
+MLX_CMAKE_ARGS=
+endif
+
+$(LIB): sources/vllm.cpp $(MLX_STAMP)
 	mkdir -p build && \
 	cd build && \
-	cmake ../sources/vllm.cpp $(CMAKE_ARGS) && \
+	cmake ../sources/vllm.cpp $(CMAKE_ARGS) $(MLX_CMAKE_ARGS) && \
 	cmake --build . --config Release -j$(JOBS) --target vllm_shared
 	cp -fL build/$(LIB) ./$(LIB)
 
@@ -79,12 +144,12 @@ vllm-cpp: main.go govllmcpp.go backend.go options.go $(LIB)
 	CGO_ENABLED=0 $(GOCMD) build -tags "$(GO_TAGS)" -o vllm-cpp ./
 
 package: vllm-cpp
-	bash package.sh
+	MLX_ROOT="$(MLX_ROOT)" bash package.sh
 
 build: package
 
 clean: purge
-	rm -rf libvllm.so libvllm.dylib package sources/vllm.cpp vllm-cpp
+	rm -rf libvllm.so libvllm.dylib package sources/vllm.cpp vllm-cpp "$(MLX_VENV)"
 
 purge:
 	rm -rf build

backend/go/vllm-cpp/README.md

@@ -41,5 +41,50 @@ options:
 - max_num_seqs:16
 ```
 
+## Apple Silicon: the MLX GEMM provider (ON by default, gated to prefill)
+
+`BUILD_TYPE=metal` builds vllm.cpp's MLX provider for the dense GEMM
+(`VLLM_CPP_MLX=on`, the default here). It is on because upstream now SHAPE-GATES
+it to prefill; it was briefly off in this branch's history, and that was correct
+at the time for an ungated provider.
+
+The gate matters more than the flag. MLX's steel GEMM wins prefill but loses
+decode, because the provider pays an `mx::eval` synchronisation plus an output
+memcpy on every call and decode makes ~112 calls *per token*. Measured on an
+Apple M4, Qwen3-1.7B-bf16 warm at p=512 g=128:
+
+| configuration | prefill TTFT | warm throughput |
+|---|--:|--:|
+| MLX **gated to prefill** (pin >= 89c46aeb) | **524.5 ms** | **24.37 tok/s, 97.6% of MLX-LM** |
+| MLX ungated (older pins) | 537 ms | 12.7 tok/s |
+| MLX off | 602 ms | 23.9 tok/s, 95.9% |
+
+Ratios are against an MLX-LM baseline measured INTERLEAVED with ours over four
+ABBA blocks (its spread 0.34%, ours 0.12%). An earlier revision of this file
+claimed 99.1%; that used a two-run MLX-LM baseline containing an outlier and
+overstated us by about 1.5 points.
+
+**`VLLM_CPP_VERSION` and this flag are coupled.** Moving the pin back before
+`89c46aeb` while leaving `VLLM_CPP_MLX=on` would take the middle row — roughly
+half throughput. If you roll the pin back, roll the default back with it.
+
+One caveat: MLX's GEMM is not bit-identical to the native kernel, so an MLX build
+produces a different greedy sequence than a non-MLX one. That is a property of the
+provider, not of the gate, and it predates this packaging. Full disposition in
+vllm.cpp `docs/BENCHMARKS.md`.
+
+Build knobs:
+
+- `VLLM_CPP_MLX=off` builds Metal without the provider: ~124 MB smaller, and
+  96.4% of MLX-LM instead of 99.1%.
+- `MLX_VERSION` pins the wheel (default `0.29.4`). MLX is consumed as the
+  prebuilt pip wheel because building it from source needs `xcrun metal`, i.e. a
+  full Xcode the macOS runners do not have.
+
+Packaging vendors `libmlx.dylib`, `mlx.metallib` and MLX's MIT license into
+`package/lib/`, and rewrites `libvllm.dylib`'s rpath to `@loader_path/lib`
+(re-signing it, since `install_name_tool` invalidates the signature). The
+metallib must stay beside `libmlx.dylib`: MLX looks for it there.
+
 Testing: `make test` runs the unit specs; export `VLLM_CPP_MODEL=<model>` (and
 optionally `VLLM_CPP_LIBRARY=<libvllm path>`) to enable the e2e specs.

backend/go/vllm-cpp/package.sh

@@ -43,6 +43,50 @@ elif [ -f "/lib/ld-linux-aarch64.so.1" ]; then
     cp -arfLv /lib/aarch64-linux-gnu/libpthread.so.0 $CURDIR/package/lib/libpthread.so.0
 elif [ $(uname -s) = "Darwin" ]; then
     echo "Detected Darwin"
+    # Vendor the optional MLX GEMM provider, when libvllm was built against it.
+    # Three facts drive every line below, each verified on an Apple M4 before it
+    # was written:
+    #   1. libvllm.dylib carries an LC_LOAD_DYLIB on @rpath/libmlx.dylib, and its
+    #      build-time LC_RPATH points inside the build venv. That path does not
+    #      exist on a user's machine, so it must become @loader_path/lib.
+    #   2. MLX finds its ~100 MB mlx.metallib beside its OWN dylib, so the two
+    #      files have to land in the same directory or every Metal op dies with
+    #      "Failed to load the default metallib".
+    #   3. install_name_tool invalidates the code signature, and macOS refuses to
+    #      load an arm64 image whose signature does not match, so the patched
+    #      library must be re-signed ad-hoc afterwards.
+    if otool -L "$CURDIR/package/libvllm.dylib" 2>/dev/null | grep -q "libmlx.dylib"; then
+        MLX_LIB_DIR="${MLX_ROOT}/lib"
+        if [ ! -f "$MLX_LIB_DIR/libmlx.dylib" ] || [ ! -f "$MLX_LIB_DIR/mlx.metallib" ]; then
+            echo "Error: libvllm.dylib links libmlx.dylib but $MLX_LIB_DIR is missing libmlx.dylib/mlx.metallib" >&2
+            exit 1
+        fi
+        echo "Vendoring the MLX GEMM provider from $MLX_LIB_DIR"
+        cp -fLv "$MLX_LIB_DIR/libmlx.dylib" "$CURDIR/package/lib/"
+        cp -fLv "$MLX_LIB_DIR/mlx.metallib" "$CURDIR/package/lib/"
+        # MLX is MIT and we redistribute its binaries, so its license ships with
+        # them. mlx-metal is the wheel carrying the dylib and the metallib.
+        MLX_LICENSE=$(ls "${MLX_ROOT}"/../mlx_metal-*.dist-info/licenses/LICENSE 2>/dev/null | head -1)
+        if [ -z "$MLX_LICENSE" ]; then
+            MLX_LICENSE=$(ls "${MLX_ROOT}"/../mlx-*.dist-info/licenses/LICENSE 2>/dev/null | head -1)
+        fi
+        if [ -z "$MLX_LICENSE" ]; then
+            echo "Error: could not find the MLX LICENSE to redistribute alongside libmlx.dylib" >&2
+            exit 1
+        fi
+        cp -fLv "$MLX_LICENSE" "$CURDIR/package/lib/LICENSE.mlx"
+        # Drop every build-tree rpath, then point at the packaged copy.
+        otool -l "$CURDIR/package/libvllm.dylib" | awk '/LC_RPATH/{f=1;next} f&&/ path /{print $2;f=0}' | while read -r rp; do
+            install_name_tool -delete_rpath "$rp" "$CURDIR/package/libvllm.dylib" 2>/dev/null || true
+        done
+        install_name_tool -add_rpath "@loader_path/lib" "$CURDIR/package/libvllm.dylib"
+        codesign -f -s - "$CURDIR/package/libvllm.dylib"
+        # A broken rpath must fail the BUILD, not the user's first inference.
+        if ! otool -l "$CURDIR/package/libvllm.dylib" | grep -q "@loader_path/lib"; then
+            echo "Error: libvllm.dylib did not get the @loader_path/lib rpath" >&2
+            exit 1
+        fi
+    fi
 else
     echo "Error: Could not detect architecture"
     exit 1

backend/python/fish-speech/run.sh

@@ -6,4 +6,9 @@ else
     source $backend_dir/../common/libbackend.sh
 fi
 
+cuda_home=${CUDA_HOME:-/usr/local/cuda}
+if [ -z "${TRITON_PTXAS_PATH:-}" ] && [ -x "$cuda_home/bin/ptxas" ]; then
+    export TRITON_PTXAS_PATH="$cuda_home/bin/ptxas"
+fi
+
 startBackend $@

core/http/react-ui/package-lock.json

@@ -24,7 +24,7 @@
         "@modelcontextprotocol/sdk": "^1.30.0",
         "dompurify": "^3.4.12",
         "highlight.js": "^11.11.1",
-        "hono": "4.12.25",
+        "hono": "4.12.34",
         "i18next": "^26.0.8",
         "i18next-browser-languagedetector": "^8.2.1",
         "i18next-http-backend": "^3.0.6",
@@ -636,12 +636,12 @@
       }
     },
     "node_modules/@hono/node-server": {
-      "version": "1.19.14",
-      "resolved": "https://registry.npmjs.org/@hono/node-server/-/node-server-1.19.14.tgz",
-      "integrity": "sha512-GwtvgtXxnWsucXvbQXkRgqksiH2Qed37H9xHZocE5sA3N8O8O8/8FA3uclQXxXVzc9XBZuEOMK7+r02FmSpHtw==",
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/@hono/node-server/-/node-server-2.1.0.tgz",
+      "integrity": "sha512-XovyyCCnBzW+zKu+z/zq8hwNs4KOR5rEMAOxo2f40Q5xoOI37IMm6MIg2COOUtUApo0i6850MTBKH2u4QLGIqg==",
       "license": "MIT",
       "engines": {
-        "node": ">=18.14.1"
+        "node": ">=20"
       },
       "peerDependencies": {
         "hono": "^4"
@@ -2879,9 +2879,9 @@
       "dev": true
     },
     "node_modules/fast-uri": {
-      "version": "3.1.4",
-      "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.4.tgz",
-      "integrity": "sha512-8JnbkQ4juDyvYs4mgFGQqg4yCYtFDtUtmp2QIQq11ZZe5CFQ5wcqm1rqDgAh/QdMySuBnPzMUiJUNZG5N/AiQw==",
+      "version": "3.1.5",
+      "resolved": "https://registry.npmjs.org/fast-uri/-/fast-uri-3.1.5.tgz",
+      "integrity": "sha512-gHwA1O9LDIcKunMKhObS/HimwtehO1nPUECKAu5TpKgaO19fcWEl4bliWe1jWxVFvIXztJjjQ4L8XQ1EU9f7Jw==",
       "funding": [
         {
           "type": "github",
@@ -3435,9 +3435,9 @@
       }
     },
     "node_modules/hono": {
-      "version": "4.12.25",
-      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.25.tgz",
-      "integrity": "sha512-2NFaIyNVgJmBs/ecmtGzlmluTFs5cHEWGTdu0t1HBwYzoGXOL5nUQBRMXsXWla5i4KkG//QMzVP88m1+I3fdAQ==",
+      "version": "4.12.34",
+      "resolved": "https://registry.npmjs.org/hono/-/hono-4.12.34.tgz",
+      "integrity": "sha512-GqXJqY/xJkJmuloTrnV1ZEXG3fqte+VjkUqoRNZXcrUidiUOP4fMSIHHY4tsqZBK++kVyWmt/AAfSUuy57/eSA==",
       "license": "MIT",
       "engines": {
         "node": ">=16.9.0"
@@ -4196,9 +4196,9 @@
       "integrity": "sha512-k/vGaX4/Yla3WzyMCvTQOXYeIHvqOKtnqBduzTHpzpQZzAskKMhZ2K+EnBiSM9zGSoIFeMpXKxa4dYeZIQqewQ=="
     },
     "node_modules/ip-address": {
-      "version": "10.2.0",
-      "resolved": "https://registry.npmjs.org/ip-address/-/ip-address-10.2.0.tgz",
-      "integrity": "sha512-/+S6j4E9AHvW9SWMSEY9Xfy66O5PWvVEJ08O0y5JGyEKQpojb0K0GKpz/v5HJ/G0vi3D2sjGK78119oXZeE0qA==",
+      "version": "10.4.0",
+      "resolved": "https://registry.npmjs.org/ip-address/-/ip-address-10.4.0.tgz",
+      "integrity": "sha512-oSK96Grm3aP6OrS263xVxbNDGVL7rzBtYdpGqlDG8iQdoenDoTs/nkki+DflYbAEE8Xl6o5YxhxlrKvI3nqKXQ==",
       "license": "MIT",
       "engines": {
         "node": ">= 12"
@@ -7138,9 +7138,9 @@
       }
     },
     "node_modules/undici": {
-      "version": "7.28.0",
-      "resolved": "https://registry.npmjs.org/undici/-/undici-7.28.0.tgz",
-      "integrity": "sha512-cRZYrTDwWznlnRiPjggAGxZXanty6M8RV1ff8Wm4LWXBp7/IG8v5DnOm74DtUBp9OONpK75YlPnIjQqX0dBDtA==",
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/undici/-/undici-7.29.0.tgz",
+      "integrity": "sha512-IDxfleLmmbSskfWSUATiN1nfn2rDuvnMOqb5CWR92iIfojA0Ud+ulOAAEQ57LPr9rWmsreUyf5lwyao+7GNNVw==",
       "dev": true,
       "license": "MIT",
       "engines": {

core/http/react-ui/package.json

@@ -19,7 +19,7 @@
     "coverage:report": "nyc report"
   },
   "overrides": {
-    "hono": "4.12.25"
+    "hono": "4.12.34"
   },
   "dependencies": {
     "@codemirror/autocomplete": "^6.18.6",
@@ -38,7 +38,7 @@
     "@modelcontextprotocol/sdk": "^1.30.0",
     "dompurify": "^3.4.12",
     "highlight.js": "^11.11.1",
-    "hono": "4.12.25",
+    "hono": "4.12.34",
     "i18next": "^26.0.8",
     "i18next-browser-languagedetector": "^8.2.1",
     "i18next-http-backend": "^3.0.6",

core/services/advisorylock/leader_loop_test.go

@@ -54,62 +54,57 @@ var _ = Describe("RunLeaderLoop", func() {
 				close(done)
 			}()
 
-			// Let it run a bit then cancel
-			time.Sleep(150 * time.Millisecond)
+			Eventually(func() int32 {
+				return atomic.LoadInt32(&callCount)
+			}, 500*time.Millisecond, 10*time.Millisecond).Should(BeNumerically(">=", 1))
 			cancel()
 
-			// RunLeaderLoop should return
 			Eventually(done, 500*time.Millisecond).Should(BeClosed())
-
-			// Record count after cancellation
-			countAfterCancel := atomic.LoadInt32(&callCount)
-			time.Sleep(150 * time.Millisecond)
-			countLater := atomic.LoadInt32(&callCount)
-
-			Expect(countLater).To(Equal(countAfterCancel),
-				"function should stop being called after context cancellation")
 		})
 
 		It("only one leader executes at a time (two concurrent loops)", func() {
 			db := testutil.SetupTestDB()
 			const lockKey int64 = 5002
 
-			var (
-				mu         sync.Mutex
-				maxRunning int32
-				running    int32
-			)
+			var running int32
+			entered := make(chan struct{}, 2)
+			release := make(chan struct{})
+			var releaseOnce sync.Once
 
 			ctx, cancel := context.WithCancel(context.Background())
-			defer cancel()
+			done := make(chan struct{}, 2)
+			DeferCleanup(func() {
+				cancel()
+				releaseOnce.Do(func() { close(release) })
+			})
 
 			fn := func() {
-				cur := atomic.AddInt32(&running, 1)
-				mu.Lock()
-				if cur > maxRunning {
-					maxRunning = cur
+				atomic.AddInt32(&running, 1)
+				select {
+				case entered <- struct{}{}:
+				default:
 				}
-				mu.Unlock()
-
-				time.Sleep(30 * time.Millisecond)
-
+				<-release
 				atomic.AddInt32(&running, -1)
 			}
 
-			// Start two competing leader loops with the same lock key
-			go RunLeaderLoop(ctx, db, lockKey, 50*time.Millisecond, fn)
-			go RunLeaderLoop(ctx, db, lockKey, 50*time.Millisecond, fn)
+			for range 2 {
+				go func() {
+					RunLeaderLoop(ctx, db, lockKey, 1*time.Millisecond, fn)
+					done <- struct{}{}
+				}()
+			}
+
+			Eventually(entered, 500*time.Millisecond).Should(Receive())
+			Consistently(func() int32 {
+				return atomic.LoadInt32(&running)
+			}, 50*time.Millisecond, 5*time.Millisecond).Should(Equal(int32(1)),
+				"expected only the lock holder to run while both loops tick")
 
-			// Let them run for a while
-			time.Sleep(400 * time.Millisecond)
 			cancel()
-
-			mu.Lock()
-			observed := maxRunning
-			mu.Unlock()
-
-			Expect(observed).To(BeNumerically("<=", 1),
-				"expected at most 1 goroutine running the leader function at a time")
+			releaseOnce.Do(func() { close(release) })
+			Eventually(done, 500*time.Millisecond).Should(Receive())
+			Eventually(done, 500*time.Millisecond).Should(Receive())
 		})
 	})
 })

docs/content/features/text-to-audio.md

@@ -192,6 +192,24 @@ You can use the env variable COQUI_LANGUAGE to set the language used by the coqu
 
 You can also use config files to configure tts models (see section below on how to use config files).
 
+### Fish Speech
+
+Fish Speech models accept the `compile` backend option. Enabling it can improve
+inference performance on CUDA hardware, but the first request after loading the
+model includes the `torch.compile` warmup cost:
+
+```yaml
+backend: fish-speech
+options:
+  - compile:true
+```
+
+When compilation is enabled, the backend uses the CUDA toolkit's executable
+`ptxas` from `$CUDA_HOME/bin` (defaulting to `/usr/local/cuda/bin`) instead of
+the copy bundled with Triton. This allows newer GPU architectures supported by
+the installed CUDA toolkit to compile kernels. Set `TRITON_PTXAS_PATH` on the
+backend explicitly to select a different assembler.
+
 ### Piper
 
 To install the `piper` audio models manually:

gallery/index.yaml

@@ -785,35 +785,18 @@
 - name: "qwen3.6-35b-a3b-uncensored-genesis-hermes-v6"
   url: "github:mudler/LocalAI/gallery/virtual.yaml@master"
   urls:
+    - https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
     - https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V6-GGUF
   description: |
-    # Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive
+    Qwen3.6-35B-A3B Uncensored Genesis Hermes V6 is LuffyTheFox's multimodal,
+    agentic derivative of HauhauCS's uncensored Qwen3.6-35B-A3B model. It
+    combines Genesis tensor calibration with Hermes function-calling data while
+    retaining the 35B mixture-of-experts architecture, roughly 3B active
+    parameters per token, and the native 262K-token context window.
 
-    > **Join the Discord** for updates, roadmaps, projects, or just to chat.
-
-    Qwen3.6-35B-A3B uncensored by HauhauCS. **0/465 Refusals.**
-
-    > **HuggingFace's "Hardware Compatibility" widget doesn't recognize K_P quants** — it may show fewer files than actually exist. Click **"View +X variants"** or go to **Files and versions** to see all available downloads.
-
-    ## About
-
-    No changes to datasets or capabilities. Fully functional, 100% of what the original authors intended - just without the refusals.
-
-    These are meant to be the best lossless uncensored models out there.
-
-    ## Aggressive Variant
-
-    Stronger uncensoring — model is fully unlocked and won't refuse prompts. May occasionally append short disclaimers (baked into base model training, not refusals) but full content is always generated.
-
-    For a more conservative uncensor that keeps some safety guardrails, check the Balanced variant when it's available.
-
-    ## Downloads
-
-    All quants generated with importance matrix (imatrix) for optimal quality preservation on abliterated weights.
-
-    ## What are K_P quants?
-
-    ...
+    This entry installs the Q8_0 GGUF together with its F16 multimodal projector
+    for llama.cpp. The model card recommends Jinja chat templates and at least a
+    128K context for its thinking behavior. License: Apache-2.0.
   license: "apache-2.0"
   tags:
     - llm

scripts/build/fish-speech-ptxas_test.sh

@@ -0,0 +1,39 @@
+#!/bin/bash
+set -euo pipefail
+
+WORK=$(mktemp -d)
+trap 'rm -rf "$WORK"' EXIT
+
+REPO_ROOT=$(dirname "$(dirname "$(dirname "$(realpath "$0")")")")
+BACKEND_DIR="$WORK/fish-speech"
+mkdir -p "$BACKEND_DIR/common" "$WORK/cuda/bin"
+cp "$REPO_ROOT/backend/python/fish-speech/run.sh" "$BACKEND_DIR/run.sh"
+
+cat > "$BACKEND_DIR/common/libbackend.sh" <<'LIBBACKEND'
+startBackend() {
+    printf '%s\n' "${TRITON_PTXAS_PATH:-}"
+}
+LIBBACKEND
+
+fail() {
+    echo "FAIL: $*"
+    exit 1
+}
+
+touch "$WORK/cuda/bin/ptxas"
+chmod +x "$WORK/cuda/bin/ptxas"
+
+got=$(CUDA_HOME="$WORK/cuda" bash "$BACKEND_DIR/run.sh")
+[ "$got" = "$WORK/cuda/bin/ptxas" ] || \
+    fail "expected toolkit ptxas, got '$got'"
+
+got=$(CUDA_HOME="$WORK/cuda" TRITON_PTXAS_PATH=/custom/ptxas \
+    bash "$BACKEND_DIR/run.sh")
+[ "$got" = "/custom/ptxas" ] || \
+    fail "explicit TRITON_PTXAS_PATH was overwritten with '$got'"
+
+chmod -x "$WORK/cuda/bin/ptxas"
+got=$(CUDA_HOME="$WORK/cuda" bash "$BACKEND_DIR/run.sh")
+[ -z "$got" ] || fail "non-executable ptxas was selected as '$got'"
+
+echo "PASS: fish-speech selects a usable toolkit ptxas"

scripts/build/llama-cpp-build-target_test.sh

@@ -29,4 +29,11 @@ assert_target arm64 "" llama-cpp-cpu-all
 assert_target amd64 sycl_f16 llama-cpp-fallback
 assert_target amd64 sycl_f32 llama-cpp-fallback
 
+# ROCm exhausts the same 6h budget through volume rather than a stall: hipcc
+# compiles ggml's HIP kernels once per AMDGPU target, eleven of them, and the
+# CPU variant matrix goes on top. 2h27m before it was added, killed at exactly
+# 6h00m on every run since.
+assert_target amd64 hipblas llama-cpp-fallback
+assert_target arm64 hipblas llama-cpp-fallback
+
 echo "PASS: llama.cpp build target preserves CPU variants where supported"