diff --git a/web-pages/product-site/data/deployments.json b/web-pages/product-site/data/deployments.json index 1bf319cae..7721ad123 100644 --- a/web-pages/product-site/data/deployments.json +++ b/web-pages/product-site/data/deployments.json @@ -176,28 +176,28 @@ "models": ["SenseVoiceSmall-GGUF", "Paraformer-GGUF", "Fun-ASR-Nano-GGUF", "FSMN-VAD-GGUF"], "operating_systems": ["Linux", "macOS", "Windows"], "interfaces": ["CLI", "local HTTP server"], - "tested": {"funasr": "runtime-llamacpp-v0.2.1", "runtime": "llama.cpp@803b7fca", "verified": "2026-08-27"}, + "tested": {"funasr": "runtime-llamacpp-v0.2.2", "runtime": "llama.cpp@05be4863", "verified": "2026-08-28"}, "commands": { - "install": ["curl -fLO https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-linux-x64.tar.gz", "echo \"1bc83dd36d11b1a2b9a45faaf77a5d1014000ac641c93dfec73d24d71ec51caf funasr-llamacpp-linux-x64.tar.gz\" | sha256sum -c -", "mkdir funasr-llamacpp && tar -xzf funasr-llamacpp-linux-x64.tar.gz -C funasr-llamacpp && cd funasr-llamacpp && bash download-funasr-model.sh sensevoice ./funasr-gguf f16"], + "install": ["curl -fLO https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-linux-x64.tar.gz", "echo \"c1ad11bac292288a783c1e5eb1103c6db301b98bf861d29b5b5310de3a190404 funasr-llamacpp-linux-x64.tar.gz\" | sha256sum -c -", "mkdir funasr-llamacpp && tar -xzf funasr-llamacpp-linux-x64.tar.gz -C funasr-llamacpp && cd funasr-llamacpp && bash download-funasr-model.sh sensevoice ./funasr-gguf f16"], "launch": ["cd funasr-llamacpp && ./llama-funasr-sensevoice -m funasr-gguf/sensevoice-small-f16.gguf --vad funasr-gguf/fsmn-vad.gguf -a sample.wav"], "health": ["cd funasr-llamacpp && ./llama-funasr-sensevoice --help"], "smoke": ["cd funasr-llamacpp && ./llama-funasr-sensevoice -m funasr-gguf/sensevoice-small-f16.gguf --vad funasr-gguf/fsmn-vad.gguf -a sample.wav | tee transcript.txt && test -s transcript.txt"] }, "downloads": [ - {"operating_system": "Linux", "architecture": "arm64", "backend": "CPU", "archive": "funasr-llamacpp-linux-arm64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-linux-arm64.tar.gz", "sha256": "9657e519986a5db72082aab508439559987fd4e563733d11da4662f898c75c45"}, - {"operating_system": "Linux", "architecture": "x64", "backend": "CPU", "archive": "funasr-llamacpp-linux-x64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-linux-x64.tar.gz", "sha256": "1bc83dd36d11b1a2b9a45faaf77a5d1014000ac641c93dfec73d24d71ec51caf"}, - {"operating_system": "Linux", "architecture": "x64 AVX2", "backend": "CPU", "archive": "funasr-llamacpp-linux-x64-avx2.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-linux-x64-avx2.tar.gz", "sha256": "32e459673297fc01fcda6901ca7d63cc9e64896a43a3c9edcb7a8e1ef70fcd04"}, - {"operating_system": "Linux", "architecture": "x64", "backend": "Vulkan", "archive": "funasr-llamacpp-linux-x64-vulkan.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-linux-x64-vulkan.tar.gz", "sha256": "d45d3acb77e05c4bae94d818ed4a5c66852be87b22c52723bd576377c055c6e6"}, - {"operating_system": "macOS", "architecture": "arm64", "backend": "CPU", "archive": "funasr-llamacpp-macos-arm64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-macos-arm64.tar.gz", "sha256": "bc63c4d4b96f2465f1d258600668a971f4f600d661f1859b03797cefaa417167"}, - {"operating_system": "Windows", "architecture": "x64", "backend": "CPU", "archive": "funasr-llamacpp-windows-x64.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-windows-x64.zip", "sha256": "b8f2b8f241b57921d82d64068d9b5695629779f3db5f3205a730cb3810232bb4"}, - {"operating_system": "Windows", "architecture": "x64 AVX2", "backend": "CPU", "archive": "funasr-llamacpp-windows-x64-avx2.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-windows-x64-avx2.zip", "sha256": "67e428a91614f8a0c1d53dc2d79cce8efc7916c030ae2ddce2cbcf667ec1c502"}, - {"operating_system": "Windows", "architecture": "x64", "backend": "Vulkan", "archive": "funasr-llamacpp-windows-x64-vulkan.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-windows-x64-vulkan.zip", "sha256": "5fd179c6aefc18477eb79fa6235c70d131a96eba621318bb84e901ae85035578"}, - {"operating_system": "Windows", "architecture": "x64", "backend": "CUDA", "archive": "funasr-llamacpp-windows-x64-cuda.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/funasr-llamacpp-windows-x64-cuda.zip", "sha256": "af32e0ef352880fcfd5589e0152c44038bc81f462f42c7a6ca91a0d9879bac58"} + {"operating_system": "Linux", "architecture": "arm64", "backend": "CPU", "archive": "funasr-llamacpp-linux-arm64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-linux-arm64.tar.gz", "sha256": "adc0e968d70a4308191a91011a3444e0a1cbefc90940c3e199b7d65ff9e59d0b"}, + {"operating_system": "Linux", "architecture": "x64", "backend": "CPU", "archive": "funasr-llamacpp-linux-x64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-linux-x64.tar.gz", "sha256": "c1ad11bac292288a783c1e5eb1103c6db301b98bf861d29b5b5310de3a190404"}, + {"operating_system": "Linux", "architecture": "x64 AVX2", "backend": "CPU", "archive": "funasr-llamacpp-linux-x64-avx2.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-linux-x64-avx2.tar.gz", "sha256": "cb5b5679938d2001426b5ea079ba948bac5c23b19aa2fc79e7a8572d9e9516e7"}, + {"operating_system": "Linux", "architecture": "x64", "backend": "Vulkan", "archive": "funasr-llamacpp-linux-x64-vulkan.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-linux-x64-vulkan.tar.gz", "sha256": "f865659d1787a2769d4ecfba598f2a490144819945bb2397fce3e172c1a1aff9"}, + {"operating_system": "macOS", "architecture": "arm64", "backend": "CPU", "archive": "funasr-llamacpp-macos-arm64.tar.gz", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-macos-arm64.tar.gz", "sha256": "cb90c64c6c251d9df9a40193037713feaee9dd602d59b470bb4735d78c00da33"}, + {"operating_system": "Windows", "architecture": "x64", "backend": "CPU", "archive": "funasr-llamacpp-windows-x64.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-windows-x64.zip", "sha256": "19e368fe0debaf880ae5aed063a1105e6cdc2d1a57f259b065df15db45a0103a"}, + {"operating_system": "Windows", "architecture": "x64 AVX2", "backend": "CPU", "archive": "funasr-llamacpp-windows-x64-avx2.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-windows-x64-avx2.zip", "sha256": "f1c9ba8e35c273b995877e0fd7f4080df28e40bce2c41be96d862c44e20fea53"}, + {"operating_system": "Windows", "architecture": "x64", "backend": "Vulkan", "archive": "funasr-llamacpp-windows-x64-vulkan.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-windows-x64-vulkan.zip", "sha256": "86a7d5ca7c134ae2fd3c9c1b356fcff041e29baf569c1145a75c73bb5bc5ea90"}, + {"operating_system": "Windows", "architecture": "x64", "backend": "CUDA", "archive": "funasr-llamacpp-windows-x64-cuda.zip", "url": "https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/funasr-llamacpp-windows-x64-cuda.zip", "sha256": "52d5ecf4220e428737f9954b148d7ee1410109a12ef77fd45f99bc1c7dd040d3"} ], "evidence": [ {"label": "llama.cpp runtime", "url": "https://github.com/modelscope/FunASR/blob/main/runtime/llama.cpp/README.md"}, - {"label": "runtime v0.2.1 release", "url": "https://github.com/modelscope/FunASR/releases/tag/runtime-llamacpp-v0.2.1"}, - {"label": "nine-platform release workflow", "url": "https://github.com/modelscope/FunASR/actions/runs/32991388379"}, + {"label": "runtime v0.2.2 release", "url": "https://github.com/modelscope/FunASR/releases/tag/runtime-llamacpp-v0.2.2"}, + {"label": "nine-platform release workflow", "url": "https://github.com/modelscope/FunASR/actions/runs/33182316846"}, {"label": "regression tests", "url": "https://github.com/modelscope/FunASR/tree/main/runtime/llama.cpp/tests"} ], "benchmarks": [ @@ -231,27 +231,27 @@ "translations": { "zh": { "name": "llama.cpp / GGUF 独立运行", - "summary": "使用 v0.2.1 的九个预编译包或源码构建，在 CPU、Vulkan、CUDA 和边缘设备上运行 FunASR GGUF 模型。", + "summary": "使用 v0.2.2 的九个预编译包或源码构建，在 CPU、Vulkan、CUDA 和边缘设备上运行 FunASR GGUF 模型。", "fit": ["不依赖 Python ML 环境", "桌面应用和离线边缘部署", "需要 Linux、macOS、Windows 的 CPU、Vulkan 或 CUDA 发布包"], "not_fit": ["需要 vLLM 式大批量 GPU 调度", "未验证目标 GPU 架构的通用预编译 CUDA 包", "必须使用完整 Python 模型生态的流程"], "selection_reason": "GGUF 运行时和独立二进制优先满足可移植、离线和低依赖部署。", - "primary_limitation": "v0.2.1 修复了匹配集成 GPU 的设备选择并优先匹配独显，但 AMD Radeon 780M 仍需报告者实机复测；RX 9070 XT 的 0xC0000005 初始化崩溃是单独的未解决问题，Android/Mali 也不是本版预编译或验证目标。", + "primary_limitation": "v0.2.2 增加确定性的后端空值检查与三段诊断边界，但不宣称已经修复 RX 9070 XT 等 AMD Windows 0xC0000005 崩溃；Android/Mali 也不是本版预编译或验证目标。", "status_label": "生产验证", "operations": ["按页面列出的 SHA-256 校验九个发布资产", "模型和二进制使用同一发布清单", "保留旧二进制和模型目录用于回滚"], "security": ["默认只读取本地音频和模型", "HTTP 服务绑定内网地址并限制上传大小", "不要从不可信地址加载 GGUF"], - "troubleshooting": ["CPU 路径先用 q8 模型验证", "CUDA 和 Vulkan 必须匹配本机驱动", "Windows AMD Vulkan 异常时记录 GPU、驱动、GGML_VK_MAX_NODES_PER_SUBMIT 和 GGML_VK_SERIALIZE_SUBMISSIONS 后回报"] + "troubleshooting": ["CPU 路径先用 q8 模型验证", "CUDA 和 Vulkan 必须匹配本机驱动", "按顺序记录 initializing ... backend on ...、initialized ... backend on ...; resolving buffer type 和 ... backend ready on ... 三段 stderr 边界", "Windows AMD Vulkan 异常时附上完整压缩包名、GPU、驱动、命令、退出码和三段诊断边界"] }, "en": { "name": "llama.cpp / GGUF standalone", - "summary": "Use the nine v0.2.1 release packages or source builds to run FunASR GGUF models on CPU, Vulkan, CUDA, and edge devices.", + "summary": "Use the nine v0.2.2 release packages or source builds to run FunASR GGUF models on CPU, Vulkan, CUDA, and edge devices.", "fit": ["No Python ML environment", "Desktop applications and offline edge deployment", "CPU, Vulkan, or CUDA packages for Linux, macOS, and Windows"], "not_fit": ["vLLM-style large GPU batch scheduling", "A universal prebuilt CUDA package for an unverified GPU architecture", "Workflows that require the full Python model ecosystem"], "selection_reason": "GGUF runtimes and standalone binaries prioritize portability, offline use, and low dependency count.", - "primary_limitation": "v0.2.1 fixes matching integrated-GPU selection and prefers a matching discrete GPU, but AMD Radeon 780M still needs reporter hardware retesting; the RX 9070 XT 0xC0000005 initialization crash remains a separate unresolved issue, and Android/Mali is not a prebuilt or validated target.", + "primary_limitation": "v0.2.2 adds deterministic backend null checks and three diagnostic boundaries, but does not claim to fix RX 9070 XT or every AMD Windows 0xC0000005 crash; Android/Mali is not a prebuilt or validated target.", "status_label": "Production verified", "operations": ["Verify all nine release assets against the listed SHA-256 values", "Keep models and binaries on the same release manifest", "Retain the previous binary and model directory for rollback"], "security": ["Read local audio and models by default", "Bind the HTTP server to a private address and limit uploads", "Do not load GGUF files from untrusted sources"], - "troubleshooting": ["Validate the q8 CPU path first", "Match CUDA or Vulkan to the local driver", "For Windows AMD Vulkan failures, record the GPU, driver, GGML_VK_MAX_NODES_PER_SUBMIT, and GGML_VK_SERIALIZE_SUBMISSIONS before reporting"] + "troubleshooting": ["Validate the q8 CPU path first", "Match CUDA or Vulkan to the local driver", "Capture the ordered stderr boundaries: initializing ... backend on ..., initialized ... backend on ...; resolving buffer type, and ... backend ready on ...", "For Windows AMD Vulkan failures, include the exact archive, GPU, driver, command, exit code, and last diagnostic boundary reached"] } } }, diff --git a/web-pages/product-site/tests/browser/product-site.spec.ts b/web-pages/product-site/tests/browser/product-site.spec.ts index 7fe3c4a0c..3f07c21d7 100644 --- a/web-pages/product-site/tests/browser/product-site.spec.ts +++ b/web-pages/product-site/tests/browser/product-site.spec.ts @@ -135,14 +135,15 @@ for (const viewport of [ { name: 'mobile', width: 390, height: 844 }, { name: 'desktop', width: 1440, height: 900 }, ]) { - test(`llama.cpp v0.2.1 download matrix is stable at ${viewport.name}`, async ({ page }, testInfo) => { + test(`llama.cpp v0.2.2 download matrix is stable at ${viewport.name}`, async ({ page }, testInfo) => { await page.setViewportSize(viewport); await page.goto('/deploy/llama-cpp.html'); const section = page.locator('[data-section="downloads"]'); await expect(section.locator('[data-download-asset]')).toHaveCount(9); - await expect(section.locator('a[href*="runtime-llamacpp-v0.2.1"]')).toHaveCount(9); + await expect(section.locator('a[href*="runtime-llamacpp-v0.2.2"]')).toHaveCount(9); await expect(page.getByText('Windows AMD Vulkan', { exact: false }).first()).toBeVisible(); + await expect(page.getByText('resolving buffer type', { exact: false }).first()).toBeVisible(); await section.evaluate((node) => node.scrollIntoView({ block: 'start' })); diff --git a/web-pages/product-site/tests/test_output.py b/web-pages/product-site/tests/test_output.py index a2412facd..4d2484c9d 100644 --- a/web-pages/product-site/tests/test_output.py +++ b/web-pages/product-site/tests/test_output.py @@ -158,16 +158,20 @@ def test_realtime_page_publishes_verified_v142_quickstart(built_site): ('en/deploy/llama-cpp.html', 'Windows AMD'), ), ) -def test_llama_cpp_pages_render_v021_download_matrix(built_site, relative, boundary): +def test_llama_cpp_pages_render_v022_download_matrix(built_site, relative, boundary): soup = read_soup(built_site / relative) section = soup.select_one('[data-section="downloads"]') assert section rows = section.select('[data-download-asset]') assert len(rows) == 9 - assert all(row.select_one('a[href*="runtime-llamacpp-v0.2.1"]') for row in rows) + assert all(row.select_one('a[href*="runtime-llamacpp-v0.2.2"]') for row in rows) assert all(len(row.select_one('[data-field="sha256"]').get_text(strip=True)) == 64 for row in rows) assert boundary in soup.get_text(' ', strip=True) + text = soup.get_text(' ', strip=True) + assert 'initializing' in text + assert 'resolving buffer type' in text + assert 'backend ready' in text @pytest.mark.parametrize( diff --git a/web-pages/product-site/tests/test_registry.py b/web-pages/product-site/tests/test_registry.py index fef13e0be..226a089ad 100644 --- a/web-pages/product-site/tests/test_registry.py +++ b/web-pages/product-site/tests/test_registry.py @@ -211,13 +211,13 @@ def test_sensevoice_tensorrt_contract_tracks_merged_native_runtime(valid_registr assert 'tensorrt version' in limitation -def test_llama_cpp_contract_tracks_v021_release_assets(valid_registry): +def test_llama_cpp_contract_tracks_v022_release_assets(valid_registry): entry = next(item for item in valid_registry['deployments'] if item['id'] == 'llama-cpp') assert entry['tested'] == { - 'funasr': 'runtime-llamacpp-v0.2.1', - 'runtime': 'llama.cpp@803b7fca', - 'verified': '2026-08-27', + 'funasr': 'runtime-llamacpp-v0.2.2', + 'runtime': 'llama.cpp@05be4863', + 'verified': '2026-08-28', } assert len(entry['downloads']) == 9 assert {item['archive'] for item in entry['downloads']} == { @@ -232,10 +232,10 @@ def test_llama_cpp_contract_tracks_v021_release_assets(valid_registry): 'funasr-llamacpp-windows-x64-cuda.zip', } assert all(item['url'].startswith( - 'https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.1/' + 'https://github.com/modelscope/FunASR/releases/download/runtime-llamacpp-v0.2.2/' ) for item in entry['downloads']) assert all(len(item['sha256']) == 64 for item in entry['downloads']) - assert any('actions/runs/32991388379' in item['url'] for item in entry['evidence']) + assert any('actions/runs/33182316846' in item['url'] for item in entry['evidence']) assert any( 'download-funasr-model.sh sensevoice ./funasr-gguf f16' in command for command in entry['commands']['install'] @@ -244,6 +244,10 @@ def test_llama_cpp_contract_tracks_v021_release_assets(valid_registry): assert 'AMD' in entry['translations']['en']['primary_limitation'] assert 'RX 9070 XT' in entry['translations']['en']['primary_limitation'] assert 'Android/Mali' in entry['translations']['en']['primary_limitation'] + troubleshooting = ' '.join(entry['translations']['en']['troubleshooting']) + assert 'initializing' in troubleshooting + assert 'resolving buffer type' in troubleshooting + assert 'backend ready' in troubleshooting def test_sensevoice_native_server_contract_tracks_merged_runtime(valid_registry):