Nexa SDK Files

Hexagon HTP

Microsoft-signed HTP catalog (llama.cpp @ ae9291). No cert import required on Windows on Snapdragon.

What's Changed

• docs: simplify Docker install to Docker Hub only by @zhiyuan8 in https://github.com/qualcomm/GenieX/pull/1139
• docs(dx): add same NPU troubleshooting to Chinese docs by @zhycheng614 in https://github.com/qualcomm/GenieX/pull/1142
• feat(server): add server-wide nctx/ngl/compute defaults to serve by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1141
• fix(ci): unify source license headers to BSD-3-Clause by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1144
• docs: drop HOME export step from Linux ARM64 install by @alanzhuly in https://github.com/qualcomm/GenieX/pull/1145
• fix(python): fix VLM multi-turn image guard and test message order by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1130
• fix(sdk): report qairt prefill rate over padded prompt length by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1148
• fix(sdk): load non-ASCII image paths in qairt VLM on Windows by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1152
• docs: add Simplified Chinese README and sync cn docs by @RichardDoong in https://github.com/qualcomm/GenieX/pull/1146
• chore(ci): drop dead .apk content-type branch in release.js by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1158
• docs(python): update install instructions to use production PyPI by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1151
• feat(sdk): support input_ids and get_model_info in qairt plugin by @Davidqian123 in https://github.com/qualcomm/GenieX/pull/1157
• ci(release): use Trusted Publishing for production PyPI by @zhycheng614 in https://github.com/qualcomm/GenieX/pull/1156
• feat(sdk): support batched prompts in geniex-bench via --- separator by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1159
• docs(linux): surface host-dependency setup as an install pre-step by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1167
• feat(qairt): opt-in sliding window CLI flag and env var by @Davidqian123 in https://github.com/qualcomm/GenieX/pull/1160
• ci(build-sdk): build unsigned HTP skels when signing secret is absent by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1171
• fix(cli): align serve short aliases with infer/run by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1172
• ci(build-sdk): override cached HTP cert path so unsigned fork builds skip signing by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1173
• feat(sdk): pull models from Docker Hub's ai/* namespace by @ericcurtin in https://github.com/qualcomm/GenieX/pull/1165
• fix(sdk): default to npu and stop clobbering explicit ngl by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1174
• feat(cli): fetch updates from S3 index instead of GitHub releases by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1164
• chore(deps): bump actions/setup-node from 6 to 7 by @dependabot[bot] in https://github.com/qualcomm/GenieX/pull/1176
• fix(sdk): fail cleanly on CPUs missing required armv8.2 features by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1180
• chore(sdk): bump geniex-qairt submodule pointer by @Davidqian123 in https://github.com/qualcomm/GenieX/pull/1175
• refactor(sdk): remove dead license_key / validation subsystem by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1184
• chore(sdk): bump llama.cpp to b10019 by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1182
• Remove Mixpanel integration from documentation by @alexchen4ai in https://github.com/qualcomm/GenieX/pull/1188
• chore(deps): bump actions/setup-go from 6 to 7 by @dependabot[bot] in https://github.com/qualcomm/GenieX/pull/1189
• docs(docs): expand contributing guide for open-source contributors by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1177
• docs(cli): sync Chinese README with English CLI docs by @HengNguan in https://github.com/qualcomm/GenieX/pull/1192
• fix(ci): bump Windows Hexagon SDK to 6.6.0.0 for parity with Linux/Android by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1193
• fix(python): exclude resumed bytes from download speed by @mengshengwu in https://github.com/qualcomm/GenieX/pull/1190
• fix(sdk): fix gemma-4-E2B VLM load and MTP quant selection by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1197
• fix(sdk): enable clip GPU backend when n_gpu_layers is -1 by @RemiliaForever in https://github.com/qualcomm/GenieX/pull/1198
• chore(deps): bump pypa/gh-action-pypi-publish from 1.14.0 to 1.14.1 by @dependabot[bot] in https://github.com/qualcomm/GenieX/pull/1200

New Contributors

• @RichardDoong made their first contribution in https://github.com/qualcomm/GenieX/pull/1146
• @ericcurtin made their first contribution in https://github.com/qualcomm/GenieX/pull/1165
• @HengNguan made their first contribution in https://github.com/qualcomm/GenieX/pull/1192

Full Changelog: https://github.com/qualcomm/GenieX/compare/v0.3.14...v0.3.16