Pinned issues
Issues
- 2
- 4
[Feature] How to Apply SAHI to training?
#1412 opened by thiagoribeirodamotta - 0
[Feature] Support sequentially loading for Large Model
#1337 opened by LZHgrla - 0
- 1
[Feature] Utilize LSP to support auto-completion, type-hinting and even code navigation.
#1398 opened by Asthestarsfalll - 2
[Bug] runner.train() forcefully initializes model weights when the model has attribute function .init_weights()
#1335 opened by Qiulin-W - 2
- 2
- 2
[Feature] Default Sampler in Runner
#1391 opened by KyanChen - 2
[Feature] Run multiple validation sets during training
#1406 opened by shani-sony - 1
How to change dataset settings and add optimizer to parameters other than model parameters when using Runner?
#1414 opened by haochuan-li - 4
- 2
[Bug] OneCycleLR object has no attribute 'end',(use before super().__init__())
#1401 opened by whlook - 9
[Feature] Support OmegaConf-style resolvers in configs
#1358 opened by collinmccarthy - 2
[Feature] Loss abstraction
#1399 opened by AwePhD - 1
[Bug] Using Deepspeed: the meaning of inputs_to_half=[0]
#1405 opened by KyanChen - 0
- 2
[Feature Request] Support passing instantiated objects to runner instead of dict configs
#1375 opened by hiyyg - 1
[Bug] xtuner wandb
#1342 opened by vansin - 1
[Feature] More Details about Module Not Registry
#1392 opened by KyanChen - 2
[Feature] Resume Validation Automatically
#1394 opened by collinmccarthy - 2
[Feature] Support LambdaLR scheduler
#1376 opened by hiyyg - 0
- 0
- 1
[Bug]
#1372 opened by yangff0016 - 0
[Feature]
#1367 opened by Aoldcat1997 - 0
[Docs] yapf 0.40.2升级造成新安装mmengine存在问题
#1364 opened by vicnoah - 2
[Feature] Customize the log directory name
#1346 opened by harimkang - 1
[Bug] cache_randomness 🐞bug
#1343 opened by HarborYuan - 0
[Feature] Implement gradient checkpointing
#1258 opened by zhouzaida - 1
[Feature] ConcatDataset metainfo check supports numpy array
#1330 opened by NrealLzx - 1
- 1
[Feature] Request for supporting load_state_dict of a sharped model when zeros have been enable
#1273 opened by Li-Qingyun - 2
[Bug] build_optim_wrapper raise an Error
#1323 opened by zero0kiriyu - 2
i want to add custom dataset exported from Roboflow using semantic segmentation masks format
#1316 opened by saeedadeeb103 - 0
[Feature] Provide an example for multitask learning
#1317 opened by zhouzaida - 0
[Bug] Multiprocessing fails to launch on Ascend NPU device when import mmengine.dist
#1307 opened by xuuyangg - 1
[Feature] about config
#1301 opened by caj-github - 3
[Feature] Can we only save a part of model parameters?
#1294 opened by hmtbgc - 4
[Bug] Multi-gpu training
#1248 opened by zbl929 - 2
- 3
[Bug] cannot build custom transforms in custom Dataset
#1287 opened by YuEn-git - 0
- 0
[Feature] Fix Visuliazer that built `vis_backends` will not be used when `save_dir` is None
#1267 opened by HAOCHENYE - 0
OpenMMLab CodeCamp 2023, Apply NOW!
#1269 opened by zhouzaida - 0
- 2
[Bug] fsdp `_move_module_to_device()` will not change the special variable `device` of **Module** `DetDataPreprocessor` to GPU!!!
#1244 opened by mypydl - 2
- 0
[Feature] Filter the frozen `Parameter` in `DefaultOptimWrapperConstructor`
#1245 opened by HAOCHENYE - 0
[Feature] Add a flag "fast_validate_before_train" to enable a fast validation before starting training
#1238 opened by zhouzaida