[pre-commit] black version

.github/workflows/lint.yml

@@ -7,14 +7,12 @@ on:
   pull_request:
 
 jobs:
-  # TODO(roller): uncomment this. it drifted due to click versioning.
-  # see #4481 for details
-  # pre-commit:
-  #   runs-on: ubuntu-latest
-  #   steps:
-  #   - uses: actions/checkout@v2
-  #   - uses: actions/setup-python@v2
-  #   - uses: pre-commit/action@v2.0.0
+  pre-commit:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - uses: actions/setup-python@v2
+    - uses: pre-commit/action@v2.0.0
 
   lint:
     runs-on: ubuntu-latest

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
 -   repo: https://github.com/psf/black
-    rev: 19.3b0
+    rev: 22.3.0
     hooks:
     -   id: black
         language_version: python3

parlai/agents/hugging_face/t5.py

@@ -345,7 +345,7 @@ def output(self, tensor):
         # Taken directly from HuggingFace
         # Rescale output before projecting on vocab
         # See https://github.com/tensorflow/mesh/blob/fa19d69eafc9a482aff0b59ddd96b025c0cb207d/mesh_tensorflow/transformer/transformer.py#L586
-        tensor = tensor * (self.t5.model_dim ** -0.5)
+        tensor = tensor * (self.t5.model_dim**-0.5)
         lm_logits = self.t5.lm_head(tensor)
         return lm_logits
 

parlai/agents/rag/indexers.py

@@ -211,7 +211,7 @@ def index_data(self, tensors: List[torch.Tensor]):
                 'HNSW index needs to index all data at once, results will be unpredictable otherwise.'
             )
         phi = 0
-        norms = (data ** 2).sum(dim=1)
+        norms = (data**2).sum(dim=1)
         max_norms = norms.max().item()
         phi = max(phi, max_norms)
         logging.info(f'HNSWF DotProduct -> L2 space phi={phi}')

parlai/agents/rag/modules.py

@@ -578,6 +578,6 @@ def reorder_decoder_incremental_state(
 
     @set_device
     def decoder_output(self, latent: torch.Tensor):
-        tensor = latent * (self.t5.model_dim ** -0.5)
+        tensor = latent * (self.t5.model_dim**-0.5)
         logits = self.t5.lm_head(tensor)
         return logits

parlai/agents/transformer/modules/decoder.py

@@ -270,7 +270,7 @@ def _default(val, default):
             )
         else:
             nn.init.normal_(
-                self.position_embeddings.weight, 0, self.embedding_size ** -0.5
+                self.position_embeddings.weight, 0, self.embedding_size**-0.5
             )
 
         # build the model

parlai/agents/transformer/modules/encoder.py

@@ -191,7 +191,7 @@ def _default(val, default):
             self.embeddings = nn.Embedding(
                 vocabulary_size, self.embedding_size, padding_idx=padding_idx
             )
-            nn.init.normal_(self.embeddings.weight, 0, self.embedding_size ** -0.5)
+            nn.init.normal_(self.embeddings.weight, 0, self.embedding_size**-0.5)
 
         # create the positional embeddings
         self.position_embeddings = nn.Embedding(self.n_positions, self.embedding_size)
@@ -203,7 +203,7 @@ def _default(val, default):
             )
         else:
             nn.init.normal_(
-                self.position_embeddings.weight, 0, self.embedding_size ** -0.5
+                self.position_embeddings.weight, 0, self.embedding_size**-0.5
             )
 
         # embedding normalization
@@ -220,7 +220,7 @@ def _default(val, default):
 
         if self.n_segments >= 1:
             self.segment_embeddings = nn.Embedding(self.n_segments, self.dim)
-            nn.init.normal_(self.segment_embeddings.weight, 0, self.dim ** -0.5)
+            nn.init.normal_(self.segment_embeddings.weight, 0, self.dim**-0.5)
 
         # build the model
         self.layers = self.build_layers()

parlai/agents/transformer/modules/functions.py

@@ -36,7 +36,7 @@ def create_embeddings(dictionary, embedding_size, padding_idx):
     Create and initialize word embeddings.
     """
     e = nn.Embedding(len(dictionary), embedding_size, padding_idx)
-    nn.init.normal_(e.weight, mean=0, std=embedding_size ** -0.5)
+    nn.init.normal_(e.weight, mean=0, std=embedding_size**-0.5)
     nn.init.constant_(e.weight[padding_idx], 0)
     return e
 

parlai/agents/transformer/polyencoder.py

@@ -404,7 +404,7 @@ def _get_embeddings(self, dict_, null_idx, embedding_size):
         embeddings = torch.nn.Embedding(
             len(dict_), embedding_size, padding_idx=null_idx
         )
-        torch.nn.init.normal_(embeddings.weight, 0, embedding_size ** -0.5)
+        torch.nn.init.normal_(embeddings.weight, 0, embedding_size**-0.5)
         return embeddings
 
     def attend(self, attention_layer, queries, keys, values, mask):

parlai/core/build_data.py

@@ -175,7 +175,7 @@ def download(url, path, fname, redownload=False, num_retries=5):
     download = not PathManager.exists(outfile) or redownload
     logging.info(f"Downloading {url} to {outfile}")
     retry = num_retries
-    exp_backoff = [2 ** r for r in reversed(range(retry))]
+    exp_backoff = [2**r for r in reversed(range(retry))]
 
     pbar = tqdm.tqdm(unit='B', unit_scale=True, desc='Downloading {}'.format(fname))
 

parlai/core/torch_classifier_agent.py

@@ -231,7 +231,7 @@ def update_raw(
         assert self._class_name == class_name
         assert len(true_labels) == len(pos_probs)
 
-        TO_INT_FACTOR = 10 ** self._max_bucket_dec_places
+        TO_INT_FACTOR = 10**self._max_bucket_dec_places
         # add the upper and lower bound of the values
         for label, prob in zip(true_labels, pos_probs):
             # calculate the upper and lower bound of the values

parlai/crowdsourcing/tasks/turn_annotations_static/analysis/compile_results.py

@@ -513,10 +513,10 @@ def compute_fleiss_kappa(
                 except Exception:
                     n_ij = 0.0
                 p_j[j] += n_ij
-                P_bar_sum_term += n_ij ** 2
+                P_bar_sum_term += n_ij**2
 
         p_j = [tmp / (N * number_of_raters) for tmp in p_j]
-        P_e_bar = sum([tmp ** 2 for tmp in p_j])
+        P_e_bar = sum([tmp**2 for tmp in p_j])
 
         P_bar = (P_bar_sum_term - N * number_of_raters) / (
             N * number_of_raters * (number_of_raters - 1)

parlai/tasks/casino/agents.py

@@ -95,17 +95,13 @@ def _setup_data(self, data_path):
             episode = copy.deepcopy(dialogue)
             episode[
                 'perspective'
-            ] = (
-                'mturk_agent_1'
-            )  # id of the agent whose perspective will be used in this dialog
+            ] = 'mturk_agent_1'  # id of the agent whose perspective will be used in this dialog
             episodes.append(episode)
 
             episode = copy.deepcopy(dialogue)
             episode[
                 'perspective'
-            ] = (
-                'mturk_agent_2'
-            )  # id of the agent whose perspective will be used in this dialog
+            ] = 'mturk_agent_2'  # id of the agent whose perspective will be used in this dialog
             episodes.append(episode)
 
         self.episodes = episodes

parlai/tasks/multiwoz_v22/agents.py

@@ -315,9 +315,7 @@ def setup_episodes(self, fold):
                 if raw_episode["dialogue_id"] != self.opt["dialogue_id"]:
                     continue
 
-            skip = (
-                False
-            )  # need to skip outer for loop while in `for domains` inner for loop
+            skip = False  # need to skip outer for loop while in `for domains` inner for loop
             if self.opt.get("well_formatted_domains_only", True):
                 if len(domains) == 0:
                     skip = True

parlai/utils/bpe.py

@@ -608,10 +608,10 @@ def bytes_to_unicode(self) -> Dict[int, str]:
         )
         cs: List[int] = bs[:]
         n = 0
-        for b in range(2 ** 8):
+        for b in range(2**8):
             if b not in bs:
                 bs.append(b)
-                cs.append(2 ** 8 + n)
+                cs.append(2**8 + n)
                 n += 1
         str_cs: List[str] = [chr(n) for n in cs]
         return dict(zip(bs, str_cs))

parlai/utils/distributed.py

@@ -212,7 +212,7 @@ def sync_parameters(model: torch.nn.Module) -> bool:
             dist.all_reduce(p.data, dist.ReduceOp.SUM)
 
     # double check everything synced correctly
-    norm2 = sum((p.data ** 2).sum().float().item() for p in model.parameters())
+    norm2 = sum((p.data**2).sum().float().item() for p in model.parameters())
     all_versions = all_gather_list(norm2)
     if not all(n == norm2 for n in all_versions):
         raise AssertionError(

parlai/utils/fp16.py

@@ -122,8 +122,8 @@ def __init__(self, optimizer, aggregate_gnorms=False):
             raise NotImplementedError("Need to implement the parameter group transfer.")
         optimizer.param_groups[0]['params'] = self.fp32_params
 
-        self.scaler = DynamicLossScaler(2.0 ** 15)
-        self.min_loss_scale = 2 ** -5
+        self.scaler = DynamicLossScaler(2.0**15)
+        self.min_loss_scale = 2**-5
         self._aggregate_gnorms = aggregate_gnorms
 
     @classmethod
@@ -318,7 +318,7 @@ class DynamicLossScaler(object):
 
     def __init__(
         self,
-        init_scale: float = 2.0 ** 15,
+        init_scale: float = 2.0**15,
         scale_factor: float = 2.0,
         scale_window: int = 2000,
         tolerance: float = 0.00,
@@ -415,7 +415,7 @@ def __init__(
         self,
         init_optimizer: torch.optim.Optimizer,  # type: ignore
         aggregate_gnorms: bool = False,
-        loss_initial_scale: float = 2.0 ** 17,
+        loss_initial_scale: float = 2.0**17,
         min_loss_scale: float = 1e-4,
     ):
         self.optimizer = init_optimizer
@@ -832,7 +832,7 @@ def step(self, closure=None):
                 group['lr'] = self._get_lr(group, state)
 
                 beta2t = 1.0 - math.pow(state['step'], group['decay_rate'])
-                update = (grad ** 2) + group['eps'][0]
+                update = (grad**2) + group['eps'][0]
                 if factored:
                     exp_avg_sq_row = state['exp_avg_sq_row']
                     exp_avg_sq_col = state['exp_avg_sq_col']

projects/image_chat/transresnet_multimodal/modules.py

@@ -526,7 +526,7 @@ def __init__(
                 n_positions, hidden_dim, out=self.position_embeddings.weight
             )
         else:
-            nn.init.normal_(self.position_embeddings.weight, 0, hidden_dim ** -0.5)
+            nn.init.normal_(self.position_embeddings.weight, 0, hidden_dim**-0.5)
 
         self.layers = nn.ModuleList()
         for _ in range(self.n_layers):

projects/light_whoami/agents/expanded_attention.py

@@ -57,7 +57,7 @@
 
 
 def get_classifier_model_and_dict(
-    opt: Opt
+    opt: Opt,
 ) -> Tuple[Optional[TorchAgent], Optional[DictionaryAgent]]:
     """
     Build classifier model and dictionary.
@@ -707,9 +707,14 @@ def _apply_model_parallel_with_extra(
         new_incr_state = {i: [] for i, _ in enumerate(self.layers)}
 
         for chunk_idx, layer_nos, next_device in work_items:
-            s_tensor, s_enc_out, s_enc_mask, s_incr_state, s_extra_out, s_extra_mask = chunks[
-                chunk_idx
-            ]
+            (
+                s_tensor,
+                s_enc_out,
+                s_enc_mask,
+                s_incr_state,
+                s_extra_out,
+                s_extra_mask,
+            ) = chunks[chunk_idx]
             for layer_no in layer_nos:
                 s_tensor, nis = self.layers[layer_no](
                     x=s_tensor,
@@ -721,7 +726,13 @@ def _apply_model_parallel_with_extra(
                 )
                 new_incr_state[layer_no].append(nis)
             # don't move incr state, it's always on the correct device
-            s_tensor, s_enc_out, s_enc_mask, s_extra_out, s_extra_mask = PipelineHelper.chunk_to(
+            (
+                s_tensor,
+                s_enc_out,
+                s_enc_mask,
+                s_extra_out,
+                s_extra_mask,
+            ) = PipelineHelper.chunk_to(
                 (s_tensor, s_enc_out, s_enc_mask, s_extra_out, s_extra_mask),
                 next_device,
             )

projects/safety_bench/model_wrappers/example_wrapper.py

@@ -29,6 +29,4 @@ def get_response(self, input_text: str) -> str:
         # Be sure to reset the model's dialogue history before/after
         # every call to `get_response`.
 
-        return (
-            "Hello"
-        )  # In this example, we always respond 'Hello' regardless of the input
+        return "Hello"  # In this example, we always respond 'Hello' regardless of the input

projects/seeker/agents/seeker.py

@@ -831,9 +831,11 @@ def batch_act(self, observations: List[Dict[str, Message]]) -> List[Message]:
         """
         knowledge_agent_observations = [o['knowledge_agent'] for o in observations]
         # First, determine whether we're searching
-        batch_reply_sdm, search_indices, knowledge_agent_observations = self.batch_act_sdm(
-            observations, knowledge_agent_observations
-        )
+        (
+            batch_reply_sdm,
+            search_indices,
+            knowledge_agent_observations,
+        ) = self.batch_act_sdm(observations, knowledge_agent_observations)
         # Second, generate search queries
         batch_reply_sqm = self.batch_act_sqm(observations, search_indices)
 

projects/seeker/agents/seeker_modules.py

@@ -243,9 +243,13 @@ def encoder(
 
         assert all(t is None for t in [input_turns_cnt, positions, segments])
         # Encode with `super()` call for non-skip-retrieval inputs
-        enc_out_retrieval, mask_retrieval, input_turns_cnt, top_docs, top_doc_scores = super(
-            ComboFidModel, self
-        ).encoder(
+        (
+            enc_out_retrieval,
+            mask_retrieval,
+            input_turns_cnt,
+            top_docs,
+            top_doc_scores,
+        ) = super(ComboFidModel, self).encoder(
             input[~skip_retrieval_vec],
             input_lengths[~skip_retrieval_vec],
             query_vec[~skip_retrieval_vec],
@@ -258,7 +262,12 @@ def encoder(
             input[skip_retrieval_vec]
         )
 
-        new_out, new_mask, new_top_docs, new_top_doc_scores = interleave_fid_combo_outputs(
+        (
+            new_out,
+            new_mask,
+            new_top_docs,
+            new_top_doc_scores,
+        ) = interleave_fid_combo_outputs(
             enc_out_retrieval,
             enc_out_skip_retrieval,
             mask_retrieval,

projects/wizard_of_wikipedia/generator/agents.py

@@ -105,7 +105,7 @@ def _set_text_vec(self, obs, history, truncate):
 class EndToEndAgent(_GenericWizardAgent):
     def __init__(self, opt, shared=None):
         super().__init__(opt, shared)
-        self._vectorize_text = lru_cache(int(2 ** 20))(self._vectorize_text)
+        self._vectorize_text = lru_cache(int(2**20))(self._vectorize_text)
 
         # knowledge truncate defaults to the same as --truncate
         self.knowledge_truncate = opt.get('knowledge_truncate')

tests/crowdsourcing/tasks/acute_eval/test_acute_eval.py

@@ -58,7 +58,6 @@ def test_base_task(
             # Check that the agent state is as it should be
             self._test_agent_state(task_data=task_data, data_regression=data_regression)
 
-
 except ImportError:
     pass
 

tests/crowdsourcing/tasks/acute_eval/test_fast_acute_dataset.py

@@ -87,7 +87,6 @@ def setup_teardown(self):
                 # Tear down temp file
                 shutil.rmtree(root_dir)
 
-
 except ImportError:
     pass
 

tests/crowdsourcing/tasks/acute_eval/test_fast_acute_logs.py

@@ -103,7 +103,6 @@ def setup_teardown(self):
                 # Tear down temp file
                 shutil.rmtree(root_dir)
 
-
 except ImportError:
     pass
 

tests/crowdsourcing/tasks/acute_eval/test_fast_acute_self_chat.py

@@ -79,7 +79,6 @@ def setup_teardown(self):
                 # Tear down temp file
                 shutil.rmtree(root_dir)
 
-
 except ImportError:
     pass
 

tests/crowdsourcing/tasks/model_chat/test_image_stack.py

@@ -72,6 +72,5 @@ def test_fill_stack(self, file_regression: FileRegressionFixture):
                 # Check the output against what it should be
                 file_regression.check(contents=stdout)
 
-
 except ImportError:
     pass

tests/crowdsourcing/tasks/model_chat/test_model_chat.py

@@ -157,7 +157,6 @@ def _remove_non_deterministic_keys(self, actual_state: dict) -> dict:
 
             return actual_state
 
-
 except ImportError:
     pass
 

tests/crowdsourcing/tasks/model_chat/test_model_chat_analysis.py

@@ -132,6 +132,5 @@ def test_worker_results_file(
                 outputs = setup_teardown
                 file_regression.check(outputs[prefix], basename=prefix)
 
-
 except ImportError:
     pass