Danling

Commits
- 4fb4023: fix lints (Zhiyuan Chen)
- e8582f0: [pre-commit.ci] pre-commit autoupdate (20) (pre-commit-ci[bot]) [20](https://github.com/ZhiyuanChen/DanLing/pull/20)
- 6f58d08: fix bug im MLP (Zhiyuan Chen)
- b25814a: fix metrics with list output (Zhiyuan Chen)
- 2312ad2: fix run_id is not consistent given the same state (Zhiyuan Chen)
- 2b05a63: [pre-commit.ci] pre-commit autoupdate (21) (pre-commit-ci[bot]) [21](https://github.com/ZhiyuanChen/DanLing/pull/21)
- 5cf2c9d: fix typo (Zhiyuan Chen)
- e2640fa: fix metrics in ddp training (Zhiyuan Chen)
- 0ee720e: [pre-commit.ci] pre-commit autoupdate (22) (pre-commit-ci[bot]) [22](https://github.com/ZhiyuanChen/DanLing/pull/22)
- 666a0cd: lazy import torchmetrics (Zhiyuan Chen)
- d72b734: deprecate python 3.7 (Zhiyuan Chen)
- 888109b: update dependencies (Zhiyuan Chen)
- 08dde7e: update doc site description (Zhiyuan Chen)
- fc2ac73: fix doc of catch decorator (Zhiyuan Chen)
- 005b395: add examples for ensure_dir (Zhiyuan Chen)
- c0762cc: improve docs for method_cache (Zhiyuan Chen)
- e7be604: move name/id/uuid to BaseRunner (Zhiyuan Chen)

latest
Commits
- f4d0488: move name/id/uuid to BaseRunner (Zhiyuan Chen)

Commits
- d039f7c: minor improvements (Zhiyuan Chen)
- 286b173: [pre-commit.ci] pre-commit autoupdate (16) (pre-commit-ci[bot]) [16](https://github.com/ZhiyuanChen/DanLing/pull/16)
- fda110f: fixes PNTensor.__deepcopy__ error in rare cases (Zhiyuan Chen)
- eba6aa6: fix name conflicts in __setattr__ (Zhiyuan Chen)
- 1e4f766: [pre-commit.ci] pre-commit autoupdate (17) (pre-commit-ci[bot]) [17](https://github.com/ZhiyuanChen/DanLing/pull/17)
- 1230ebf: [pre-commit.ci] pre-commit autoupdate (18) (pre-commit-ci[bot]) [18](https://github.com/ZhiyuanChen/DanLing/pull/18)
- 6cd8366: fix mypy (Zhiyuan Chen)
- cbb1c63: fix NestedTensor.storage is not a method and _state is not a property (Zhiyuan Chen)
- a48b14e: fix lint (Zhiyuan Chen)
- 1f6b5af: fix train tensorboard (Zhiyuan Chen)
- 34e5719: set default value of final_lr_ratio to 1e-3 (Zhiyuan Chen)
- e7b1fdd: make MLP and Dense more flexible (Zhiyuan Chen)
- cbeee3f: fix device for Metrics (Zhiyuan Chen)
- 7ebcf84: explicitly set kwargs for method_cache (Zhiyuan Chen)
- a074ee7: support 'pandas' and 'pd' for pandas pickle in IO (Zhiyuan Chen)
- 8a25769: update runner id (Zhiyuan Chen)
- a23ec75: make check_dir a standlone method (Zhiyuan Chen)
- 210c2fb: fix padding values used in metrics calc when input/target is NestedTensor (Zhiyuan Chen)
- 8828a4d: improve NestedTensor (Zhiyuan Chen)
- 541bfc2: improve Metrics & AverageMeters (Zhiyuan Chen)

Commits
- 02741ba: register collate_pn_tensor_fn to default_collate_fn_map (Zhiyuan Chen)
- bc30250: NestedTensor now returns False when comparing to objects with unsupported type (Zhiyuan Chen)
- edac28a: copy class config to local at init (Zhiyuan Chen)
- 7097dd1: no longer print at the first iter, but at the last iter (Zhiyuan Chen)
- 22efbb2: fix eval dataset in mnist demo (Zhiyuan Chen)

Commits
- dfa2bb1: fixes NestedTensor in metrics (Zhiyuan Chen)
- bacfbc2: [pre-commit.ci] pre-commit autoupdate (12) (pre-commit-ci[bot]) [12](https://github.com/ZhiyuanChen/DanLing/pull/12)
- 3d7a712: [pre-commit.ci] pre-commit autoupdate (13) (pre-commit-ci[bot]) [13](https://github.com/ZhiyuanChen/DanLing/pull/13)
- 1104e3d: rename steps to total_steps in LRScheduler (Zhiyuan Chen)
- feb0d0b: [pre-commit.ci] pre-commit autoupdate (14) (pre-commit-ci[bot]) [14](https://github.com/ZhiyuanChen/DanLing/pull/14)
- 54031e5: refactor result system (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 636886a: support train/evaluate/inference (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 6c884c4: fix is_best raises KeyError when results is empty (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 98621d0: supoort dict as input data (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- c4f71c2: fix epoch log when epoch is 0 (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 5652be7: merge average_meters (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 6864e3f: fix docs of metrics (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- b22e909: support autocast in default train_epoch (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2d3ab9c: minor improvements (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- c6166ee: merge RunnerBase into BaseRunner (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 1d5c67a: support __post_init__ (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 4f57415: auto prepares objects (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2c3b3f5: deprecate pass args & kwargs to build runner (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2b6fda9: improve deepspeed support (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 85f15ec: reorganise code (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 1a8c27a: fixes compatibilities with previous python (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 5b346d7: fixes requirements (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 145148c: add index in best_result/latest_result (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2e2406d: auto create dataloader (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- ef04898: fix print in step_log (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2315e91: remove patience from state, make it optional (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 8c2fc5d: fix writer and logging (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- beaa5d2: train/eval now accepts a list of splits (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 982ac83: update deepspeed config preprocess (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 171cdb7: support NestedTensor in Metrics (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 4e6722f: remove danling.registry (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 6d39448: add metrics.functional (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 4b07c73: fixes results (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 2ba08e2: rename IndexMetrics to ScoreMetrics (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 277f14b: move loss backward to step (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- e7329fa: report result to nni (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 1149708: fixes conflicting names (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 23102d3: fixes results (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- eb17e49: fixes batch_size (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 18a3f2b: fixes deepspeed init (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 96e1d30: fixes docs (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- f20b19c: no longer auto prepare LRScheduler (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- e6f7c2b: move step implementation to TorchRunner (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 420e1b8: fix RecursionError in rare cases (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 5bbf968: only init datasets/dataloaders/datasamplers when they are not inited (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 43e60db: add total_epochs/total_steps (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- 1931f29: revert b159da789d8f360a36cf91801a6f2c9ae621a697 (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- f4c8154: Runner.deepspeed now returns None if DS is not enabled (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)
- f825685: read env virable for deepspeed training (Zhiyuan Chen) [15](https://github.com/ZhiyuanChen/DanLing/pull/15)

Commits
- b24fc76: update pre-commit (Zhiyuan Chen)
- 7b16d6e: merge average_meters (Zhiyuan Chen)
- 81fbc03: make functional a standalone file (Zhiyuan Chen)
- 2237079: register collate_pn_tensor_fn to default_collate_fn_map (Zhiyuan Chen)
- 6937343: NestedTensor now returns False when comparing to objects with unsupported type (Zhiyuan Chen)
- 5887b14: copy class config to local at init (Zhiyuan Chen)
- 1b458b9: fixes PNTensor.__deepcopy__ error in rare cases (Zhiyuan Chen)
- b31dad9: fix name conflicts in __setattr__ (Zhiyuan Chen)
- 88aa731: fix NestedTensor.storage is not a method and _state is not a property (Zhiyuan Chen)
- c9c3e14: set default value of final_lr_ratio to 1e-3 (Zhiyuan Chen)
- 848cde6: make MLP and Dense more flexible (Zhiyuan Chen)
- dda7269: explicitly set kwargs for method_cache (Zhiyuan Chen)
- 6745a34: support 'pandas' and 'pd' for pandas pickle in IO (Zhiyuan Chen)
- f4b540a: fix padding values used in metrics calc when input/target is NestedTensor (Zhiyuan Chen)
- cfedf8e: improve NestedTensor (Zhiyuan Chen)
- 4ab86f4: improve Metrics & AverageMeters (Zhiyuan Chen)
- 46787c5: fix lints (Zhiyuan Chen)

Commits
- dfa2bb1: fixes NestedTensor in metrics (Zhiyuan Chen)
- bacfbc2: [pre-commit.ci] pre-commit autoupdate (12) (pre-commit-ci[bot]) [12](https://github.com/ZhiyuanChen/DanLing/pull/12)
- 3d7a712: [pre-commit.ci] pre-commit autoupdate (13) (pre-commit-ci[bot]) [13](https://github.com/ZhiyuanChen/DanLing/pull/13)
- 1104e3d: rename steps to total_steps in LRScheduler (Zhiyuan Chen)
- feb0d0b: [pre-commit.ci] pre-commit autoupdate (14) (pre-commit-ci[bot]) [14](https://github.com/ZhiyuanChen/DanLing/pull/14)