[CI] fix some spelling errors (#3707)

* fix spelling error with examples/comminity/

* fix spelling error with tests/

* fix some spelling error with tests/ colossalai/ etc.

tests/test_utils/test_activation_checkpointing.py

@@ -51,7 +51,7 @@ def test_activation_checkpointing(cpu_offload, use_reentrant):
     # other tests might affect this test
     reset_seeds()
 
-    # We put initilization here to avoid change cuda rng state below
+    # We put initialization here to avoid change cuda rng state below
     inputs = torch.rand(2, 2, requires_grad=True, device='cuda')
     weight = torch.rand(2, 4, requires_grad=True, device='cuda')
 

tests/test_utils/test_checkpoint_io/test_load.py

@@ -23,7 +23,7 @@ def check_model_state_dict(a: Dict[str, Tensor], b: Dict[str, Tensor]) -> None:
         assert torch.equal(v, b[k])
 
 
-def check_optim_state_dict(a: dict, b: dict, ignore_param_gruops: bool = False) -> None:
+def check_optim_state_dict(a: dict, b: dict, ignore_param_groups: bool = False) -> None:
     assert set(a['state'].keys()) == set(b['state'].keys())
     for k, state in a['state'].items():
         b_state = b['state'][k]
@@ -32,7 +32,7 @@ def check_optim_state_dict(a: dict, b: dict, ignore_param_gruops: bool = False)
                 assert torch.equal(v1, v2)
             else:
                 assert v1 == v2
-    if not ignore_param_gruops:
+    if not ignore_param_groups:
         assert a['param_groups'] == b['param_groups']
 
 
@@ -129,23 +129,23 @@ def launch_dist(fn, world_size: int):
 
 
 def save_dist(dir_name: str, zero: bool):
-    model, optmizer = prepare_model_optim(shard=True, zero=zero)
-    reset_model_optim(model, optmizer)
+    model, optimizer = prepare_model_optim(shard=True, zero=zero)
+    reset_model_optim(model, optimizer)
     world_size = dist.get_world_size()
     rank = dist.get_rank()
-    save(dir_name, model, optmizer, dist_meta=get_dist_metas(world_size, zero)[rank])
+    save(dir_name, model, optimizer, dist_meta=get_dist_metas(world_size, zero)[rank])
 
 
 def load_and_check_dist(dir_name: str):
     world_size = dist.get_world_size()
-    model, optmizer = prepare_model_optim(shard=True)
-    reset_model_optim(model, optmizer)
+    model, optimizer = prepare_model_optim(shard=True)
+    reset_model_optim(model, optimizer)
     model_state_dict = deepcopy(model.state_dict())
-    optimizer_state_dict = deepcopy(optmizer.state_dict())
-    reset_model_optim(model, optmizer, 1)
-    load(dir_name, model, optmizer, get_redist_meta(world_size), get_dist_metas(world_size))
+    optimizer_state_dict = deepcopy(optimizer.state_dict())
+    reset_model_optim(model, optimizer, 1)
+    load(dir_name, model, optimizer, get_redist_meta(world_size), get_dist_metas(world_size))
     check_model_state_dict(model_state_dict, model.state_dict())
-    check_optim_state_dict(optimizer_state_dict, optmizer.state_dict())
+    check_optim_state_dict(optimizer_state_dict, optimizer.state_dict())
 
 
 @pytest.mark.dist

tests/test_utils/test_checkpoint_io/test_merge.py

@@ -68,7 +68,7 @@ def run_dist(rank, world_size, port, test_fn):
 
 
 def run_save_dist(dir_name: str, zero: bool):
-    model, optmizer = prepare_model_optim(shard=True, zero=zero)
+    model, optimizer = prepare_model_optim(shard=True, zero=zero)
     rank = dist.get_rank()
     dp_world_size = dist.get_world_size() // 2
     if not zero:
@@ -90,7 +90,7 @@ def run_save_dist(dir_name: str, zero: bool):
             'fc.bias':
                 ParamDistMeta(rank // 2, dp_world_size, 0, 1, zero_numel=1, zero_orig_shape=[1])
         }
-    save(dir_name, model, optmizer, dist_meta=dist_metas)
+    save(dir_name, model, optimizer, dist_meta=dist_metas)
 
 
 @pytest.mark.dist

tests/test_utils/test_checkpoint_io/test_redist.py

@@ -125,9 +125,9 @@ def run_dist(rank, world_size, port, test_fn):
 
 
 def run_save_dist(dir_name: str, zero: bool):
-    model, optmizer = prepare_model_optim(shard=True, zero=zero)
+    model, optimizer = prepare_model_optim(shard=True, zero=zero)
     rank = dist.get_rank()
-    save(dir_name, model, optmizer, dist_meta=get_dist_metas(4, zero)[rank])
+    save(dir_name, model, optimizer, dist_meta=get_dist_metas(4, zero)[rank])
 
 
 @pytest.mark.dist

tests/test_utils/test_checkpoint_io/test_save.py

@@ -28,7 +28,7 @@ def check_model_state_dict(a: Dict[str, Tensor], b: Dict[str, Tensor]) -> None:
         assert torch.equal(v, b[k])
 
 
-def check_optim_state_dict(a: dict, b: dict, ignore_param_gruops: bool = False) -> None:
+def check_optim_state_dict(a: dict, b: dict, ignore_param_groups: bool = False) -> None:
     assert set(a['state'].keys()) == set(b['state'].keys())
     for k, state in a['state'].items():
         b_state = b['state'][k]
@@ -37,7 +37,7 @@ def check_optim_state_dict(a: dict, b: dict, ignore_param_gruops: bool = False)
                 assert torch.equal(v1, v2)
             else:
                 assert v1 == v2
-    if not ignore_param_gruops:
+    if not ignore_param_groups:
         assert a['param_groups'] == b['param_groups']
 
 
@@ -113,12 +113,12 @@ def run_dist(rank, world_size, port, test_fn):
 
 
 def run_save_dist(dir_name):
-    model, optmizer = prepare_model_optim()
+    model, optimizer = prepare_model_optim()
     dist_metas = {
         'fc.weight': ParamDistMeta(dist.get_rank(), dist.get_world_size(), 0, 1),
         'fc.bias': ParamDistMeta(dist.get_rank(), dist.get_world_size(), 0, 1)
     }
-    save(dir_name, model, optmizer, dist_meta=dist_metas)
+    save(dir_name, model, optimizer, dist_meta=dist_metas)
 
 
 @pytest.mark.dist

tests/test_utils/test_lazy_init/utils.py

@@ -18,7 +18,7 @@ def set_seed(seed: int) -> None:
     torch.manual_seed(seed)
 
 
-def assert_model_eqaual(m1: torch.nn.Module, m2: torch.nn.Module) -> None:
+def assert_model_equal(m1: torch.nn.Module, m2: torch.nn.Module) -> None:
     s1 = m1.state_dict()
     s2 = m2.state_dict()
 
@@ -63,7 +63,7 @@ def check_lazy_init(entry: TestingEntry, seed: int = 42, verbose: bool = False,
     with ctx:
         deferred_model = model_fn()
     deferred_model = ctx.materialize(deferred_model, verbose=verbose)
-    assert_model_eqaual(model, deferred_model)
+    assert_model_equal(model, deferred_model)
     if check_forward:
         assert_forward_equal(model, deferred_model, data_gen_fn, output_transform_fn)
     if verbose: