From efe23ffc2c1e6b414ef2a1631c9ed8f7f7197bd3 Mon Sep 17 00:00:00 2001
From: Jintao Huang <hjt_study@qq.com>
Date: Sat, 15 Oct 2022 19:05:20 +0800
Subject: [PATCH] v0.1.4

---
 README.md                                                | 2 +-
 examples/dqn.py                                          | 2 +-
 mini_lightning/__init__.py                               | 8 ++++----
 mini_lightning/{mini_lightning.py => _mini_lightning.py} | 2 +-
 mini_lightning/{utils.py => _utils.py}                   | 0
 mini_lightning/{visualize.py => _visualize.py}           | 0
 mini_lightning/{warmup_lrs.py => _warmup_lrs.py}         | 1 -
 7 files changed, 7 insertions(+), 8 deletions(-)
 rename mini_lightning/{mini_lightning.py => _mini_lightning.py} (99%)
 rename mini_lightning/{utils.py => _utils.py} (100%)
 rename mini_lightning/{visualize.py => _visualize.py} (100%)
 rename mini_lightning/{warmup_lrs.py => _warmup_lrs.py} (97%)

diff --git a/README.md b/README.md
index 9caf58b..3673adc 100644
--- a/README.md
+++ b/README.md
@@ -14,7 +14,7 @@
 2. Download the latest version (>=1.12) of Torch(corresponding CUDA version) from the [official website](https://pytorch.org/get-started/locally/) of Torch. It is not recommended to automatically install Torch (CUDA 10.2 default) using the Mini-Lightning dependency, which will cause CUDA version mismatch.
 3. Install mini-lightning
 ```bash
-# from pypi (v0.1.3)
+# from pypi (v0.1.4)
 pip install mini-lightning
 
 # Or download the files from the repository to local,
diff --git a/examples/dqn.py b/examples/dqn.py
index 403133f..33e37ea 100644
--- a/examples/dqn.py
+++ b/examples/dqn.py
@@ -206,7 +206,7 @@ def training_step(self, batch: Any) -> Tensor:
         "optim_name": "SGD",
         "dataloader_hparams": {"batch_size": batch_size},
         "optim_hparams": {"lr": 1e-2, "weight_decay": 1e-4},  #
-        "trainer_hparams": {"max_epochs": max_epochs, "gradient_clip_norm": 20},
+        "trainer_hparams": {"max_epochs": max_epochs, "gradient_clip_norm": 20, "verbose": False},
         #
         "rand_p": {
             "eta_max": 1,
diff --git a/mini_lightning/__init__.py b/mini_lightning/__init__.py
index d2b667b..c39ee47 100644
--- a/mini_lightning/__init__.py
+++ b/mini_lightning/__init__.py
@@ -1,4 +1,4 @@
-from .mini_lightning import *
-from .utils import *
-from .warmup_lrs import *
-from .visualize import *
+from ._mini_lightning import *
+from ._utils import *
+from ._warmup_lrs import *
+from ._visualize import *
diff --git a/mini_lightning/mini_lightning.py b/mini_lightning/_mini_lightning.py
similarity index 99%
rename from mini_lightning/mini_lightning.py
rename to mini_lightning/_mini_lightning.py
index e7b1755..d71974a 100644
--- a/mini_lightning/mini_lightning.py
+++ b/mini_lightning/_mini_lightning.py
@@ -873,7 +873,7 @@ def test(self, dataloader: Optional[DataLoader], test_best: bool = False, test_l
         #
         if test_last:  # just current model
             if self._best_ckpt_is_last() and test_best is True:
-                logger.info("Ignore test last: the best ckpt is the last ckpt")
+                logger.info("Ignore test last: the best ckpt and the last ckpt is the same")
             else:
                 m = self._test(dataloader, "last")
                 res_mes.update(m)
diff --git a/mini_lightning/utils.py b/mini_lightning/_utils.py
similarity index 100%
rename from mini_lightning/utils.py
rename to mini_lightning/_utils.py
diff --git a/mini_lightning/visualize.py b/mini_lightning/_visualize.py
similarity index 100%
rename from mini_lightning/visualize.py
rename to mini_lightning/_visualize.py
diff --git a/mini_lightning/warmup_lrs.py b/mini_lightning/_warmup_lrs.py
similarity index 97%
rename from mini_lightning/warmup_lrs.py
rename to mini_lightning/_warmup_lrs.py
index 74058b4..b70c262 100644
--- a/mini_lightning/warmup_lrs.py
+++ b/mini_lightning/_warmup_lrs.py
@@ -5,7 +5,6 @@
 import math
 from typing import List, Callable, Union, Dict, Optional
 #
-from torch.optim.lr_scheduler import _LRScheduler as LRScheduler, CosineAnnealingLR
 from torch.optim import Optimizer
 __all__ = ["get_T_max", "warmup_decorator", "cosine_annealing_lr"]