ameva-runtime 2.7.2__tar.gz → 2.7.3__tar.gz

ameva_runtime-2.7.3/PKG-INFO

@@ -0,0 +1,217 @@
+Metadata-Version: 2.4
+Name: ameva-runtime
+Version: 2.7.3
+Summary: Unified Next-Gen Hardware Orchestration & AI Acceleration Runtime for Mobile & Edge
+Home-page: https://github.com/uno-km/ameva-runtime
+Author: Eunho Kim
+Author-email: Eunho Kim <contact@uno-km.com>
+License: Apache-2.0
+Project-URL: Homepage, https://uno-km.vercel.app/lib/vulkan/
+Project-URL: Repository, https://github.com/uno-km/ameva-runtime
+Project-URL: Documentation, https://uno-km.vercel.app/lib/vulkan/
+Keywords: vulkan-compute,mobile-gpu,hardware-acceleration,hardware-abstraction-layer,adreno-gpu,arm-mali,snapdragon-8-elite,exynos,termux,on-device-ai,edge-ai,tensor-acceleration,spir-v,compute-shaders,zero-silent-fallback,llamacpp,whisper-cpp,sherpa-onnx,stable-diffusion,vision-language-models,bitnet,gguf,ncnn,bionic-loader,arm64,aarch64,cgroup-management,cpu-neon,thermal-throttling,power-efficiency,smart-router,hardware-orchestration,multi-modal-ai,subgroup-operations,gemm-acceleration,mobile-vlm,speech-to-text,text-to-speech,image-generation,edge-inference,unprivileged-userspace,termux-wake-lock,phantom-process-killer,android-ai-runtime,valhall-gpu,adreno-830,mali-g68,ameva-foundation,uno-km,open-source-ai
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: POSIX :: Linux
+Classifier: Operating System :: Android
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: stt
+Requires-Dist: termux-stt>=1.2.0; extra == "stt"
+Provides-Extra: diffusion
+Requires-Dist: termux-diffusion>=1.5.0; extra == "diffusion"
+Provides-Extra: bitnet
+Requires-Dist: termux-bitnet>=1.2.0; extra == "bitnet"
+Provides-Extra: llamacpp
+Requires-Dist: termux-llamacpp>=1.3.0; extra == "llamacpp"
+Provides-Extra: tts
+Requires-Dist: termux-tts>=1.4.0; extra == "tts"
+Provides-Extra: vision
+Requires-Dist: termux-vision>=1.2.0; extra == "vision"
+Provides-Extra: all
+Requires-Dist: termux-stt>=1.2.0; extra == "all"
+Requires-Dist: termux-diffusion>=1.5.0; extra == "all"
+Requires-Dist: termux-bitnet>=1.2.0; extra == "all"
+Requires-Dist: termux-llamacpp>=1.3.0; extra == "all"
+Requires-Dist: termux-tts>=1.4.0; extra == "all"
+Requires-Dist: termux-vision>=1.2.0; extra == "all"
+Dynamic: author
+Dynamic: home-page
+Dynamic: license-file
+Dynamic: requires-python
+
+# AMEVA-Runtime
+
+[![PyPI](https://img.shields.io/pypi/v/ameva-runtime.svg?style=flat-square&color=0369a1)](https://pypi.org/project/ameva-runtime/)
+[![Python](https://img.shields.io/pypi/pyversions/ameva-runtime.svg?style=flat-square)](https://pypi.org/project/ameva-runtime/)
+[![npm](https://img.shields.io/npm/v/@ameva/runtime.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/@ameva/runtime)
+[![GitHub Release](https://img.shields.io/github/v/release/uno-km/ameva-runtime-releases?style=flat-square&color=0969da)](https://github.com/uno-km/ameva-runtime-releases/releases/tag/v2.7.3)
+[![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/ameva-runtime)
+<img src="https://img.shields.io/badge/BitNet%201.58b-Vulkan%20Compute%20Accelerated-purple.svg?logo=vulkan&logoColor=white" alt="BitNet Vulkan">
+
+> Next-Gen Unified On-Device Hardware Orchestration & 6-Modality AI Acceleration Runtime (with BitNet 1.58-bit Vulkan Compute) for Mobile & Edge
+
+---
+
+## Architecture & Overview
+
+AMEVA-Runtime is a hardware abstraction layer (HAL) and compute orchestration engine engineered specifically for mobile ARM64 devices (Android Termux, Linux Edge). It continuously inspects underlying silicon topology (/dev/kgsl-3d0, /dev/mali0) to route tensor execution across Qualcomm Adreno, ARM Mali, and ARM Cortex CPU-NEON backends.
+
+### 6-Modality Acceleration Matrix
+
+| Modality | Engine Integration | Status (v2.7.3) | Hardware Acceleration Mechanism |
+| :--- | :--- | :---: | :--- |
+| **1. LLM (Text)** | Llama.cpp & Termux-BitNet (1.58-bit i2_s) | **Production (v2.7.3)** | Vulkan 25/25 layer VRAM offload (`ngl=999`), BitNet full pipeline & strict `AmbiguousModelMatchError` |
+| **2. STT (Speech)** | Whisper.cpp (Large-v3-Turbo) | **Production (v2.7.3 / STT v1.2.7)** | Vulkan GPU acceleration with greedy decoding default (`-bs 1`), 6-SoC fleet validated |
+| **3. TTS (Audio)** | MeloTTS / Piper / Kokoro / Supertonic | **Production (v2.7.3 / TTS v1.5.0)** | Bionic Direct Vulkan (`/system/lib64/libvulkan.so`) & HiFi-GAN 32MB buffer temporal tiling |
+| **4. Diffusion (Image)** | SDXS / SD1.5 / Z-Image Turbo (6.0B DiT) | **Production (v2.7.3 / Diffusion v1.8.0)** | Z-Image Turbo 6.0B DiT, Flash Attention (`--diffusion-fa`), VAE Tiling, Zero-Collision Bionic Vulkan backend |
+| **5. Vision (VLM)** | CLIP / MobileVLM / LLaVA | **In Development** | GGML Vulkan vision encoder tensor bindings |
+| **6. Train (Training)** | On-Device LoRA / QLoRA | **In Development** | Mobile Vulkan gradient descent backpropagation |
+
+---
+
+## Empirical Physical Device Benchmarks
+
+Tested on physical devices running Android 16 under Termux ARM64:
+
+### 1. LLM Generation (Qwen2.5-0.5B-Instruct Q4_K_M)
+| Target Device | Hardware Architecture | Active Backend | Layers in VRAM | Generation Speed | Prompt Processing | Speedup |
+| :--- | :--- | :---: | :---: | :---: | :---: | :---: |
+| **Galaxy S25** | Snapdragon 8 Elite / Adreno 830 | Vulkan 1.3 | **25/25 (100%)** | **35.80 t/s** (27.9 ms/t) | 4.53 t/s | **35.8x (vs CPU)** |
+| **Galaxy A35** | Exynos 1380 / ARM Mali-G68 MP5 | Vulkan 1.3 | **25/25 (100%)** | **4.44 t/s** (225 ms/t) | 6.12 t/s | **+26.9% (vs NEON)** |
+| **Galaxy A35** | Cortex-A78 CPU-NEON (3 Threads) | CPU-NEON | 0/25 | 3.55 t/s (281 ms/t) | 8.05 t/s | Baseline |
+
+### 2. Speech-to-Text (Whisper Large-v3-Turbo Q5_0, 548MB)
+| Target Device | Hardware Architecture | Backend Mode | Latency (1-min audio) | GPU Load | CPU Load | Speedup |
+| :--- | :--- | :---: | :---: | :---: | :---: | :---: |
+| **Galaxy S25** | Snapdragon 8 Elite / Adreno 830 | Vulkan GPU | **4.40 s (0.07x RTF)** | Adreno Turbo | ~12% | **18.5x (vs CPU)** |
+| **Galaxy S22** | Snapdragon 8 Gen 1 / Adreno 730 | Vulkan GPU | **5.13 s (Encoder)** | Adreno Native | ~14% | **3.73x (vs CPU)** |
+| **Galaxy A35** | Exynos 1380 / Mali-G68 MP5 | Vulkan GPU | **360.60 s (6m 00s)** | **949 MHz (100%)** | 20~30% | **2.26x (56% time saved)** |
+| **Galaxy A35** | Cortex-A78 x4 Cores | CPU-NEON | 816.48 s (13m 36s) | 0% | 291% | Baseline |
+
+### 3. Text-to-Speech (AMEVA Bionic Native Vulkan Fleet Benchmarks)
+| Target Device | Hardware Architecture | Neural Engine / Model | Backend Mode | Latency | RTF | Forensics (RMS/Peak) | Status |
+| :--- | :--- | :--- | :---: | :---: | :---: | :---: | :---: |
+| **Galaxy S22** | Snapdragon 8 Gen 1 / Adreno 730 | MeloTTS Universal Bilingual | **Bionic Vulkan GPU** | **2,750 ms** | **0.88x** | 0.0814 / 0.6974 | Real-time Synthesized |
+| **Galaxy S21** | Exynos 2100 / ARM Mali-G78 | Piper VITS (On-chip Tiled) | **Bionic Vulkan GPU** | **560 ms** | **0.18x** | 0.0921 / 0.7412 | 5.5x Faster than RT |
+| **Galaxy S20** | Exynos 990 / ARM Mali-G77 | Piper VITS (On-chip Tiled) | **Bionic Vulkan GPU** | **750 ms** | **0.24x** | 0.0890 / 0.7105 | 4.1x Faster than RT |
+| **Galaxy S25** | Snapdragon 8 Elite / Adreno 830 | Supertonic 3 Flow / Piper | **Bionic Vulkan GPU** | **380 ms** | **0.12x** | 0.1042 / 0.8120 | Studio Ultra-Fast |
+| **Galaxy A35** | Exynos 1380 / ARM Mali-G68 MP5 | Piper VITS (`lessac-medium`) | **Vulkan GPU** | **5,180 ms** | **1.146x** | 0.0782 / 0.6540 | Validated |
+
+---
+
+### 4. BitNet 1.58-bit LLM (Microsoft BitNet b1.58 2B-4T i2_s)
+| Target Device | Hardware Architecture | Active Backend | Generation Speed | Prompt Latency | Speedup | Status |
+| :--- | :--- | :---: | :---: | :---: | :---: | :---: |
+| **Galaxy S25** | Snapdragon 8 Elite / Adreno 830 | **AMEVA Vulkan GPU** | **17.558 t/s** | **205.9 ms** | **12.58x** | **Verified (Ground Truth)** |
+| Galaxy S25 | Snapdragon 8 Elite / Oryon CPU | Native CPU (4 Threads) | 1.396 t/s | 2,041.0 ms | 1.00x | Baseline |
+| **Galaxy A35** | Exynos 1380 / Mali-G68 MP5 | **AMEVA Vulkan GPU** | **3.471 t/s** | **1,552.8 ms** | **5.94x** | **Verified (Ground Truth)** |
+| Galaxy A35 | Exynos 1380 / Cortex-A78 CPU | Native CPU (4 Threads) | 0.584 t/s | 8,775.0 ms | 1.00x | Baseline |
+
+## Root-Cause Driver Solutions
+
+1. **ARM Mali-G68 Valhall Integer Truncation**: Enforced medium tile matmul kernel dispatch (`loadstride_b = 4 > 0`), permanently eliminating shader zero-stride infinite loops on subgroup-16 hardware.
+2. **Qualcomm Adreno 830 JIT Register Bug**: Bounded vector column specialization (`mul_mat_vec_max_cols = 2`), preventing compiler crash `VK_ERROR_UNKNOWN (-13)`.
+3. **Bionic Direct Vulkan Linking**: Permanently bypasses Termux `$PREFIX/lib/libvulkan.so` Mesa `llvmpipe` CPU software emulator trap by dynamically binding `/system/lib64/libvulkan.so` directly, routing compute shader execution to physical Adreno/Mali silicon.
+4. **Mobile GPU 32MB Memory Ceiling & Temporal Tiling**: Resolved MeloTTS HiFi-GAN 52.4MB buffer overflow (`VK_ERROR_OUT_OF_DEVICE_MEMORY` / kernel TDR kill) via temporal chunk slicing ($T_{\text{chunk}} \le 819$) and Piper VITS on-chip SRAM tiling.
+5. **Zero-Silent-Fallback & Strict Model Resolution**: Guaranteed fail-fast architecture without silent CPU degradation. Explicitly raises `AmbiguousModelMatchError` on multi-model ambiguity and rejects silent guessing.
+
+---
+
+## Two-Track Supply Chain Architecture
+
+To simultaneously guarantee **100% intellectual property & source code protection** and **frictionless, tokenless binary deployment**, AMEVA enforces a strict Two-Track Supply Chain Architecture:
+
+```
+[Track 1: Local Ground Truth & Offline Staging]
+├── dev/ameva/ameva-runtime/releases/ (Consolidated 6-Modality SOTA Binaries)
+└── dev/termux/termux-*/releases/     (Domain-Specific Compiled Binaries)
+                     │
+                     ▼ Synchronized via Git & Automated Release Workflows
+[Track 2: Remote Public GitHub Releases Cloud & CDN]
+├── uno-km/ameva-runtime-releases    (Public Unified Binary Hub: Releases v2.7.3)
+└── uno-km/termux-* (diffusion, stt) (Domain-Specific Public GitHub Releases)
+```
+
+1. **Core Source Code (`uno-km/ameva-runtime`)**: Kept **100% Private**. Architectural proprietary logic, HAL implementations, and optimization kernels are permanently protected.
+2. **Public Binary Distribution Hub (`uno-km/ameva-runtime-releases`)**: **Public Registry**. Hosts verified precompiled binaries, hardware acceleration compute shaders, and wheels for anonymous 1-Click installation without requiring personal GitHub tokens.
+
+---
+
+## Installation & Engine Auto-Provisioning
+
+### 1. Python SDK Installation
+```bash
+# Direct install from public release wheel
+pip install --upgrade "https://github.com/uno-km/ameva-runtime-releases/releases/latest/download/ameva_runtime-2.7.3-py3-none-any.whl"
+
+# Or standard PyPI install
+pip install ameva-runtime
+```
+
+### 2. 1-Click Native Engine Provisioning
+The `NativeAssetManager` automatically retrieves cryptographically verified native engine binaries from the public registry and establishes system links in `$PREFIX/bin`:
+
+```bash
+# Provision all hardware engines at once
+ameva install --all
+
+# Or provision specific engine modalities individually
+ameva install --modality diffusion   # Stable Diffusion Vulkan Engine (sd-cli)
+ameva install --modality stt         # Whisper STT Vulkan Engine (whisper-cli)
+ameva install --modality llm         # LLaMA C++ Vulkan Engine (llama-cli)
+ameva install --modality tts         # Sherpa-NCNN Neural Voice Engine
+ameva install --modality bitnet      # BitNet 1.58-bit Acceleration Engine
+
+# Validate system readiness with 12-stage hardware diagnostic
+ameva doctor
+```
+
+---
+
+## Quickstart
+
+### Python SDK
+```python
+import ameva_runtime as ameva
+from ameva_runtime import vulkan
+
+# 1. Inspect on-device silicon topology
+profile = ameva.detect_hardware()
+print(f"SoC: {profile.soc_name} | GPU: {profile.gpu_vendor}")
+
+# 2. Run hardware self-test
+doc = vulkan.Doctor()
+report = doc.run_self_test()
+print(f"GPU: {report.device_name} (Passed: {report.passed_stages}/{report.total_stages})")
+```
+
+### Node.js / TypeScript
+```typescript
+import { Doctor, createContext } from '@ameva/runtime';
+
+const doc = new Doctor();
+const report = await doc.runSelfTest();
+console.log(`Vulkan GPU: ${report.deviceName}`);
+```
+
+---
+
+## Official Documentation & Benchmarks
+- [Official Architecture & API Reference](https://uno-km.vercel.app/lib/vulkan/)
+- [Ecosystem Metrics & Registry Stats](https://uno-km.vercel.app/foundation/metrics)
+- [AMEVA Open-Source Foundation Portal](https://uno-km.vercel.app/foundation/index.html)
+
+---
+
+## License
+Licensed under the Apache-2.0 License. Copyright (c) 2026 Eunho Kim ([@uno-km](https://github.com/uno-km)).

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/README.md

@@ -3,7 +3,7 @@
 [![PyPI](https://img.shields.io/pypi/v/ameva-runtime.svg?style=flat-square&color=0369a1)](https://pypi.org/project/ameva-runtime/)
 [![Python](https://img.shields.io/pypi/pyversions/ameva-runtime.svg?style=flat-square)](https://pypi.org/project/ameva-runtime/)
 [![npm](https://img.shields.io/npm/v/@ameva/runtime.svg?style=flat-square&color=b91c1c)](https://www.npmjs.com/package/@ameva/runtime)
-[![GitHub Release](https://img.shields.io/github/v/release/uno-km/ameva-runtime?style=flat-square&color=0969da)](https://github.com/uno-km/ameva-runtime/releases/tag/v2.7.1)
+[![GitHub Release](https://img.shields.io/github/v/release/uno-km/ameva-runtime-releases?style=flat-square&color=0969da)](https://github.com/uno-km/ameva-runtime-releases/releases/tag/v2.7.3)
 [![License](https://img.shields.io/badge/License-Apache_2.0-004499.svg?style=flat-square)](https://github.com/uno-km/ameva-runtime)
 <img src="https://img.shields.io/badge/BitNet%201.58b-Vulkan%20Compute%20Accelerated-purple.svg?logo=vulkan&logoColor=white" alt="BitNet Vulkan">
 
@@ -17,13 +17,13 @@ AMEVA-Runtime is a hardware abstraction layer (HAL) and compute orchestration en
 
 ### 6-Modality Acceleration Matrix
 
-| Modality | Engine Integration | Status (v2.7.1) | Hardware Acceleration Mechanism |
+| Modality | Engine Integration | Status (v2.7.3) | Hardware Acceleration Mechanism |
 | :--- | :--- | :---: | :--- |
-| **1. LLM (Text)** | Llama.cpp & Termux-BitNet (1.58-bit i2_s) | **Production (v2.7.1)** | Vulkan 25/25 layer VRAM offload (`ngl=999`), BitNet full pipeline (Adreno 17.56 t/s, Mali 3.47 t/s) & strict `AmbiguousModelMatchError` |
-| **2. STT (Speech)** | Whisper.cpp (Large-v3-Turbo) | **Production (v2.7.1 / STT v1.2.7)** | Vulkan GPU acceleration with greedy decoding default (`-bs 1`), 6-SoC fleet validated (Adreno 3.73x, Mali 2.26x) |
-| **3. TTS (Audio)** | MeloTTS / Piper / Kokoro / Supertonic | **Production (v2.7.1 / TTS v1.5.0)** | Bionic Direct Vulkan (`/system/lib64/libvulkan.so`) & HiFi-GAN 32MB buffer temporal tiling (Adreno RTF 0.88x/0.264x, Mali RTF 0.18x~0.24x) |
-| **4. Vision (VLM)** | CLIP / MobileVLM / LLaVA | **In Development** | GGML Vulkan vision encoder tensor bindings |
-| **5. Diffusion (Image)** | Stable Diffusion v1.5 / FLUX.1 | **In Development** | On-device Vulkan UNet & DiT tensor offload |
+| **1. LLM (Text)** | Llama.cpp & Termux-BitNet (1.58-bit i2_s) | **Production (v2.7.3)** | Vulkan 25/25 layer VRAM offload (`ngl=999`), BitNet full pipeline & strict `AmbiguousModelMatchError` |
+| **2. STT (Speech)** | Whisper.cpp (Large-v3-Turbo) | **Production (v2.7.3 / STT v1.2.7)** | Vulkan GPU acceleration with greedy decoding default (`-bs 1`), 6-SoC fleet validated |
+| **3. TTS (Audio)** | MeloTTS / Piper / Kokoro / Supertonic | **Production (v2.7.3 / TTS v1.5.0)** | Bionic Direct Vulkan (`/system/lib64/libvulkan.so`) & HiFi-GAN 32MB buffer temporal tiling |
+| **4. Diffusion (Image)** | SDXS / SD1.5 / Z-Image Turbo (6.0B DiT) | **Production (v2.7.3 / Diffusion v1.8.0)** | Z-Image Turbo 6.0B DiT, Flash Attention (`--diffusion-fa`), VAE Tiling, Zero-Collision Bionic Vulkan backend |
+| **5. Vision (VLM)** | CLIP / MobileVLM / LLaVA | **In Development** | GGML Vulkan vision encoder tensor bindings |
 | **6. Train (Training)** | On-Device LoRA / QLoRA | **In Development** | Mobile Vulkan gradient descent backpropagation |
 
 ---
@@ -76,14 +76,53 @@ Tested on physical devices running Android 16 under Termux ARM64:
 
 ---
 
-## Installation
+## Two-Track Supply Chain Architecture
 
+To simultaneously guarantee **100% intellectual property & source code protection** and **frictionless, tokenless binary deployment**, AMEVA enforces a strict Two-Track Supply Chain Architecture:
+
+```
+[Track 1: Local Ground Truth & Offline Staging]
+├── dev/ameva/ameva-runtime/releases/ (Consolidated 6-Modality SOTA Binaries)
+└── dev/termux/termux-*/releases/     (Domain-Specific Compiled Binaries)
+                     │
+                     ▼ Synchronized via Git & Automated Release Workflows
+[Track 2: Remote Public GitHub Releases Cloud & CDN]
+├── uno-km/ameva-runtime-releases    (Public Unified Binary Hub: Releases v2.7.3)
+└── uno-km/termux-* (diffusion, stt) (Domain-Specific Public GitHub Releases)
+```
+
+1. **Core Source Code (`uno-km/ameva-runtime`)**: Kept **100% Private**. Architectural proprietary logic, HAL implementations, and optimization kernels are permanently protected.
+2. **Public Binary Distribution Hub (`uno-km/ameva-runtime-releases`)**: **Public Registry**. Hosts verified precompiled binaries, hardware acceleration compute shaders, and wheels for anonymous 1-Click installation without requiring personal GitHub tokens.
+
+---
+
+## Installation & Engine Auto-Provisioning
+
+### 1. Python SDK Installation
 ```bash
-# Python SDK
+# Direct install from public release wheel
+pip install --upgrade "https://github.com/uno-km/ameva-runtime-releases/releases/latest/download/ameva_runtime-2.7.3-py3-none-any.whl"
+
+# Or standard PyPI install
 pip install ameva-runtime
+```
+
+### 2. 1-Click Native Engine Provisioning
+The `NativeAssetManager` automatically retrieves cryptographically verified native engine binaries from the public registry and establishes system links in `$PREFIX/bin`:
 
-# Node.js / TypeScript
-npm install @ameva/runtime
+```bash
+# Provision all hardware engines at once
+ameva install --all
+
+# Or provision specific engine modalities individually
+ameva install --modality diffusion   # Stable Diffusion Vulkan Engine (sd-cli)
+ameva install --modality stt         # Whisper STT Vulkan Engine (whisper-cli)
+ameva install --modality llm         # LLaMA C++ Vulkan Engine (llama-cli)
+ameva install --modality tts         # Sherpa-NCNN Neural Voice Engine
+ameva install --modality bitnet      # BitNet 1.58-bit Acceleration Engine
+
+# Validate system readiness with 12-stage hardware diagnostic
+ameva doctor
 ```
 
 ---

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/pyproject.toml

@@ -7,7 +7,7 @@ name = "ameva-runtime"
 dynamic = ["version"]
 description = "Unified Next-Gen Hardware Orchestration & AI Acceleration Runtime for Mobile & Edge"
 
-readme = "README.pypi.md"
+readme = "README.md"
 requires-python = ">=3.8"
 license = { text = "Apache-2.0" }
 authors = [{ name = "Eunho Kim", email = "contact@uno-km.com" }]

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/_version.py

@@ -5,4 +5,4 @@ Single Source of Truth (SSOT) for all runtime packages.
 Modifying this single string updates ameva_runtime, ameva_runtime.vulkan,
 and the package metadata simultaneously.
 """
-__version__ = "2.7.2"
+__version__ = "2.7.3"

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/adapters/base.py

@@ -546,13 +546,16 @@ def get_vulkan_env(base_env: Optional[dict[str, str]] = None) -> dict[str, str]:
 
     # Mali GPU Quirks: Avoid infinite GEMM quantization loops on ARM Mali Valhall architectures
     try:
-        from ..platform import detect_soc_environment
+        from ..platform import detect_soc_environment, is_android
         soc = detect_soc_environment()
-        if "mali" in str(soc.gpu_family).lower() or soc.vendor == "samsung":
+        if "mali" in str(soc.gpu_family).lower():
             env.setdefault("GGML_VK_FORCE_MEDIUM_MATMUL", "1")
             env.setdefault("GGML_VK_DISABLE_F16", "1")
 
-            # On Samsung Exynos / ARM Mali devices, enable HAL shim to sanitize vendor queue queries and GOS symbols
+        # Mobile Vulkan HAL shim: In Android Termux environment, preload verified libegl_shim.so
+        # to intercept and resolve missing vendor Bionic symbols (e.g. _ZN7android18egl_get_connectionEv)
+        # across Samsung One UI and vendor drivers cleanly without crash.
+        if is_android() or os.path.exists("/system/lib64/libvulkan.so"):
             shim_path = resolve_vulkan_shim_path()
             if shim_path:
                 curr_preload = env.get("LD_PRELOAD", "")

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/adapters/diffusion.py

@@ -215,9 +215,9 @@ class DiffusionAdapter(BaseAdapter):
     def build_cli_args(
         cls,
         executable: str,
-        model_path: str,
-        prompt: str,
-        output_path: str,
+        model_path: Optional[str] = None,
+        prompt: str = "",
+        output_path: str = "",
         width: int = 512,
         height: int = 512,
         steps: int = 2,
@@ -231,6 +231,13 @@ class DiffusionAdapter(BaseAdapter):
         vae_tiling: bool = False,
         backend: Optional[str] = None,
         is_adreno: bool = False,
+        llm_path: Optional[str] = None,
+        diffusion_model_path: Optional[str] = None,
+        clip_l_path: Optional[str] = None,
+        flash_attn: bool = False,
+        offload_to_cpu: bool = False,
+        clip_on_cpu: bool = False,
+        vae_on_cpu: bool = False,
     ) -> list[str]:
         """최신 sd-cli 규격에 부합하는 안전하고 검증된 Diffusion CLI 인자 목록을 조립합니다."""
         if threads == "auto" or threads is None:
@@ -241,7 +248,6 @@ class DiffusionAdapter(BaseAdapter):
 
         cmd = [
             str(executable),
-            "-m", str(model_path),
             "-p", str(prompt),
             "-o", str(output_path),
             "-W", str(width),
@@ -251,16 +257,39 @@ class DiffusionAdapter(BaseAdapter):
             "--cfg-scale", str(cfg_scale),
         ]
 
+        if diffusion_model_path:
+            cmd.extend(["--diffusion-model", str(diffusion_model_path)])
+        elif model_path:
+            cmd.extend(["-m", str(model_path)])
+
+        if llm_path:
+            cmd.extend(["--llm", str(llm_path)])
+        if clip_l_path:
+            cmd.extend(["--clip_l", str(clip_l_path)])
+
         if guidance is not None:
             cmd.extend(["--guidance", str(guidance)])
 
-        if target_backend == "cpu":
-            cmd.append("--offload-to-cpu")
+        if clip_on_cpu and "--clip-on-cpu" not in cmd:
+            cmd.append("--clip-on-cpu")
+        if vae_on_cpu and "--vae-on-cpu" not in cmd:
+            cmd.append("--vae-on-cpu")
+
+        if offload_to_cpu:
+            if "--offload-to-cpu" not in cmd:
+                cmd.append("--offload-to-cpu")
+        elif target_backend == "cpu":
+            if "--offload-to-cpu" not in cmd:
+                cmd.append("--offload-to-cpu")
         elif backend:
             cmd.extend(["--backend", str(backend)])
         elif is_adreno or target_backend in ("vulkan", "gpu"):
-            cmd.extend(["--backend", "clip=vulkan0,diffusion=vulkan0,vae=vulkan0"])
+            b_clip = "cpu" if clip_on_cpu else "vulkan0"
+            b_vae = "cpu" if vae_on_cpu else "vulkan0"
+            cmd.extend(["--backend", f"clip={b_clip},diffusion=vulkan0,vae={b_vae}"])
 
+        if flash_attn:
+            cmd.append("--diffusion-fa")
         if vae_tiling:
             cmd.append("--vae-tiling")
         if sampling_method:

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/adapters/llamacpp.py

@@ -146,27 +146,10 @@ class LlamaCppAdapter(BaseAdapter):
         **kwargs: Any,
     ) -> Dict[str, str]:
         """Provides verified execution environment adhering to Golden Link Order.
-        Strictly isolates engine libraries and system Vulkan drivers without Termux symbol pollution.
+        Inherits base adapter environment (proper LD_PRELOAD for libegl_shim, Mali flags, LD_LIBRARY_PATH)
+        without Gate 1 symbol pollution or unsafe system path injection.
         """
-        env = dict(base_env or os.environ)
-        home = Path(os.environ.get("HOME", os.path.expanduser("~")))
-        std_llama_lib = home / ".local" / "share" / "ameva" / "current" / "llamacpp" / "lib"
-        legacy_llama_lib = home / ".termux-llama" / "current" / "lib"
-
-        lib_paths = []
-        if std_llama_lib.is_dir():
-            lib_paths.append(str(std_llama_lib))
-        elif legacy_llama_lib.is_dir():
-            lib_paths.append(str(legacy_llama_lib))
-
-        sys_lib64 = Path("/system/lib64")
-        if sys_lib64.is_dir():
-            lib_paths.append(str(sys_lib64))
-
-        env["LD_LIBRARY_PATH"] = ":".join(lib_paths)
-        # Ensure GGML_VULKAN_SKIP_CHECKS is never present
-        env.pop("GGML_VULKAN_SKIP_CHECKS", None)
-        return env
+        return super().get_execution_environment(base_env, **kwargs)
 
     @classmethod
     def _mutate_modality_attributes(

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/adapters/stt.py

@@ -664,8 +664,17 @@ class SttAdapter(BaseAdapter):
         base_env: dict[str, str] | None = None,
         **kwargs: Any,
     ) -> dict[str, str]:
-        """Provides verified execution environment adhering to Golden Link Order."""
-        return super().get_execution_environment(base_env, **kwargs)
+        """Provides verified execution environment adhering to Golden Link Order.
+        Strictly applies Zero-Collision Rule to purge Termux $PREFIX/lib from LD_LIBRARY_PATH,
+        preventing Android 15 Bionic linker symbol collisions (e.g. libunwindstack Xzs_Construct).
+        """
+        env = super().get_execution_environment(base_env, **kwargs)
+        if "LD_LIBRARY_PATH" in env:
+            prefix = os.environ.get("PREFIX", "/data/data/com.termux/files/usr")
+            prefix_lib = f"{prefix}/lib"
+            parts = [p for p in env["LD_LIBRARY_PATH"].split(":") if p and p != prefix_lib]
+            env["LD_LIBRARY_PATH"] = ":".join(parts)
+        return env
 
     @staticmethod
     def bind(

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/doctor.py

@@ -189,12 +189,13 @@ class Doctor:
 
     def quick_probe(self) -> bool:
         p = self.profile
-        if not p.has_vulkan_loader:
-            return False
         if p.hardware_hazard:
             return False
         if p.gpu_family in ("Unknown", "None", ""):
             return False
+        # Android Termux can use Vulkan directly via KGSL/Mali driver node even if libvulkan loader is not in $PREFIX/lib
+        if not (p.has_vulkan_loader or p.has_kgsl_node or p.has_mali_node):
+            return False
         return p.recommended_backend in ("vulkan", "vulkan_driver_only")
 
 

{ameva_runtime-2.7.2 → ameva_runtime-2.7.3}/python/ameva_runtime/installer.py

@@ -34,7 +34,7 @@ from .exceptions import (
 
 logger = logging.getLogger("ameva_runtime.installer")
 
-AMEVA_RUNTIME_GITHUB_REPO = "uno-km/ameva-runtime"
+AMEVA_RUNTIME_GITHUB_REPO = "uno-km/ameva-runtime-releases"
 GITHUB_RELEASE_LATEST = f"https://github.com/{AMEVA_RUNTIME_GITHUB_REPO}/releases/latest/download"
 
 
@@ -47,6 +47,8 @@ def get_release_base_url() -> str:
     if custom_base := os.environ.get("AMEVA_RELEASE_BASE"):
         return custom_base.rstrip("/")
     if custom_tag := os.environ.get("AMEVA_RELEASE_TAG"):
+        if custom_tag.strip().lower() == "latest":
+            return f"https://github.com/{AMEVA_RUNTIME_GITHUB_REPO}/releases/latest/download"
         tag = custom_tag if custom_tag.startswith("v") else f"v{custom_tag}"
         return f"https://github.com/{AMEVA_RUNTIME_GITHUB_REPO}/releases/download/{tag}"
     return f"https://github.com/{AMEVA_RUNTIME_GITHUB_REPO}/releases/download/v{__version__}"
@@ -306,7 +308,7 @@ NATIVE_ASSETS: Dict[str, AssetSpec] = {
     "diffusion": AssetSpec(
         name="diffusion",
         filename="sd-cli-vulkan-android-arm64.tar.gz",
-        download_sha256="0340cf0a8ca4172f184a9b5b9e419905e35b51a00e07e88841d8537224235fbc",
+        download_sha256="2d268255e84a37eca8e3ea047f79f944d23068108307f509cfd5c28484b05737",
         binary_relpath="sd-cli-vulkan",
         canonical_name="sd-cli",
         description="Stable Diffusion On-Device Vulkan CLI Engine",
@@ -665,9 +667,20 @@ class NativeAssetManager:
             try:
                 resp = urllib.request.urlopen(req, timeout=60)
             except urllib.error.HTTPError as http_err:
-                raise AmevaLlamaAssetMissingError(
-                    f"Release bundle unavailable: url={download_url}, status={http_err.code}"
-                ) from http_err
+                if http_err.code == 404 and "latest/download" not in download_url:
+                    fallback_url = f"https://github.com/{AMEVA_RUNTIME_GITHUB_REPO}/releases/latest/download/{spec.filename}"
+                    print(f"    -> [FALLBACK] Versioned tag returned 404. Attempting public latest release: {fallback_url} ...")
+                    req_fb = urllib.request.Request(fallback_url, headers={"User-Agent": f"AMEVA-Installer/{__version__}"})
+                    try:
+                        resp = urllib.request.urlopen(req_fb, timeout=60)
+                    except urllib.error.HTTPError:
+                        raise AmevaLlamaAssetMissingError(
+                            f"Release bundle unavailable: url={download_url}, status={http_err.code}"
+                        ) from http_err
+                else:
+                    raise AmevaLlamaAssetMissingError(
+                        f"Release bundle unavailable: url={download_url}, status={http_err.code}"
+                    ) from http_err
 
             downloaded = 0
             with resp, open(temp_dest, "wb") as out_f: