| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
feat(cli): integrate hyper_v2 search into run_nd CLI with demo YAMLs, parser UT, and field-name alignment (Issue 126) | 2 个月前 | |
fix: 修复静态分析告警，补充模块文档与返回值校验 | 4 个月前 | |
feat(dmodule): add M1 declarative Configurable and Module layer Introduce Configurable in hyper_parallel/config/ and declarative Module, BaseModel, ShardingConfig, and ModelSpec under hyper_parallel.dmodule. ModelSpec supports Config-based registration alongside legacy build_model_fn. Module.parallelize applies DTensor sharding on a single DeviceMesh; local_map is deferred to M9. Expand Google-style docstrings with usage examples; add resolve_placements unit tests (case sensitivity, empty mesh, Partial). Add tests/ut/dmodule and dist_module parallelize coverage without wiring the new path into LLMTrainer or package __init__ exports. | 5 个月前 | |
feat: add scheduled torch activation recomputation | 2 个月前 | |
DSV4 CSA/HCA DFunction | 3 个月前 | |
feat: support Qwen3.5 and Qwen3-VL-MoE training Add dense, MoE, and vision-language model integration with checkpoint conversion, data loading, and unified trainer execution. Correct TP, CP, EP, PP, FSDP, and HSDP synchronization semantics so supported BF16 single-card and hybrid-parallel configurations follow the same training behavior. | 3 个月前 | |
fix: clean up static code scan alerts for assigned owners Resolve H2305/C0411/W1805/H2205 findings in examples, dmodule, fully_shard, fast-tuner, and config modules; suppress C0413 where HYPER_PARALLEL_PLATFORM must be set before hyper_parallel imports. | 4 个月前 | |
Update generate checks Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> | 3 个月前 | |
| 4 个月前 | ||
feat: support Qwen3.5 and Qwen3-VL-MoE training Add dense, MoE, and vision-language model integration with checkpoint conversion, data loading, and unified trainer execution. Correct TP, CP, EP, PP, FSDP, and HSDP synchronization semantics so supported BF16 single-card and hybrid-parallel configurations follow the same training behavior. | 3 个月前 | |
feat: add scheduled torch activation recomputation | 2 个月前 | |
fix: improve memory visualizer with large file support and UI polish | 3 个月前 | |
feat: add training monitoring function | 3 个月前 | |
feat(dtensor): add from_local run_check, rand/randn factories, and public exports Align HyperParallel DTensor surface with PyTorch by exporting distribute_tensor and factory helpers from hyper_parallel, adding run_check validation for from_local, and wiring rand/randn through OffsetBasedRNGTracker. Shorten the PP+FSDP+TP 1F1B composite level1 ST from 1000 steps to 100. | 3 个月前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 2 个月前 | ||
| 4 个月前 | ||
| 5 个月前 | ||
| 2 个月前 | ||
| 3 个月前 | ||
| 3 个月前 | ||
| 4 个月前 | ||
| 3 个月前 | ||
| 4 个月前 | ||
| 3 个月前 | ||
| 2 个月前 | ||
| 3 个月前 | ||
| 3 个月前 | ||
| 3 个月前 |