phyai/examples · mingti-org/phyai

GGitHubfeat(phyai): add Dreamzero support (#65 )

文件	最后提交记录	最后更新时间
configs	feat(phyai): quantization with humming (#69)	25 天前
cosmos3	feat(phyai): add optimized staged Cosmos3-Nano inference (#73) Expose independent VLM, VAE encode, DiT, and VAE decode stages with optional offload and warmup-aware timing. Reuse CUDA graphs and optimize normalization, attention preparation, and sampling without caching DiT outputs. Move real-sample preparation into Cosmos3PolicyProcessor. Add customized inputs, bundled AgiBot assets, frame comparisons, and bilingual usage docs.	17 天前
dreamzero	feat(phyai): add Dreamzero support (#65) * feat(phyai): add DreamZero support on latest main * refactor(dreamzero): reuse framework layers and runtime TF32 settings * feat(phyai): add policy-selectable L2 RMSNorm kernel * fix(phyai): validate L2 RMSNorm and migrate text linear layers * fix(dreamzero): reuse compiled UniPC updates across timesteps Pass current and historical sigma tensors into compiled updates so complete schedules reuse bounded order variants. Preserve the original GPU BF16 trajectory. * fix(dreamzero): preserve input shapes and guidance options Normalize vector state/action samples before padding, propagate explicit negative prompts through both processor steps, and skip negative encoding and broadcasts when effective guidance disables CFG. * fix(dreamzero): correct VAE batches and encoder offload Allocate tiled decoder outputs for the full batch, convert encoder tile units once per call, and construct gated image-encoder projections on CPU before loading and activation. * refactor(dreamzero): simplify inference and add CuTe T5 GELU Remove legacy WS1 and sequential CPU offload interfaces, redundant cache wrappers, and unused encoder paths. Reuse shared tensor-parallel layers and select FlashInfer's FA2 backend for DreamZero attention. Add a JIT CUTLASS CuTe T5 GELU kernel for contiguous and strided inputs, with corresponding layer/kernel tests and benchmarks. Preserve strict checkpoint loading, numerical precision, and cache lifecycle contracts. Validation: targeted CUDA kernel/layer checks, 43 final CUDA audit cases including two-rank TP/CFG and fake checkpoint coverage, and scripts/run_pre_commit.sh. * docs(skills): generalize PR review and validation guidance Require revision-aware feedback checks, shared-layer contract verification, and complete removal of retired interfaces. Cover cache ownership, device defaults, tensor parallelism, numerical equivalence, and kernel benchmarks. Keep PR-specific evidence in work logs and preserve the user's authorization boundaries when fixing code or interacting with GitHub. * docs(dreamzero): add bilingual inference and accuracy guides * docs(dreamzero): address final review feedback --------- Co-authored-by: lywbarca <lywbarca@users.noreply.github.com> Co-authored-by: chenghuaWang <chenghua.wang.edu@gmail.com>	5 天前
gr00t	feat(phyai): refactor server and parallel group (#58) * refactor(phyai): kernel dispatch system * refactor(phyai): server and parallel group	1 个月前
minicpm_gr00t	feat(phyai): refactor server and parallel group (#58) * refactor(phyai): kernel dispatch system * refactor(phyai): server and parallel group	1 个月前
molmoact2	feat(phyai): Support molmoact2 (#70) * feat(phyai): quantization with humming * feat(phyai): support molmoact2	22 天前
moss_image	feat(phyai): moss-image (#77)	12 天前
pi0	feat(phyai): refactor server and parallel group (#58) * refactor(phyai): kernel dispatch system * refactor(phyai): server and parallel group	1 个月前
pi05	feat(phyai): quantization with humming (#69)	25 天前
qwen_image_21	feat(phyai): Qwen-Image-2.1	27 天前
xiaomi_u0	feat(phyai): add native Xiaomi U0 text-to-image inference (#71) * feat(phyai): add Xiaomi U0 reference text-to-image inference * perf(xiaomi_u0): use contiguous KV views in reference attention * chore(xiaomi_u0): prepare reference inference PR * docs(xiaomi-u0): clarify validation results and test scope * docs(xiaomi-u0): use uv for workspace setup and execution * docs(xiaomi-u0): document inference and generation contracts * feat(phyai): use native Xiaomi U0 inference and add examples * fix(docs): make U0 sample images render on GitHub * docs(xiaomi-u0): rewrite bilingual generation guides * docs(xiaomi-u0): keep validation details in the PR * fix(phyai): align U0 defaults and require explicit cache capacity * fix(phyai): gate U0 EOS on the complete image trailer * docs(xiaomi-u0): use relative paths for sample images * docs: announce Xiaomi Robotics U0 text-to-image support * docs: link Xiaomi U0 announcements to the documentation site --------- Co-authored-by: YoungHLee <941971174@qq.com>	6 天前