[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

applications/Colossal-LLaMA/colossal_llama/model/init_model.py

@@ -4,6 +4,7 @@
 """
 Initialize new model with updated tokenizer by calculating the mean values from original model
 """
+
 import argparse
 
 import numpy as np

applications/ColossalChat/coati/dataset/tokenization_utils.py

@@ -56,10 +56,8 @@ def tokenize_sft(
     template.messages = []
     for idx, mess in enumerate(messages):
         if mess["from"] != template.roles[idx % 2]:
-            raise ValueError(
-                f"Message should iterate between user and assistant and starts with a \
-                             line from the user. Got the following data:\n{messages}"
-            )
+            raise ValueError(f"Message should iterate between user and assistant and starts with a \
+                             line from the user. Got the following data:\n{messages}")
         template.append_message(mess["from"], mess["content"])
 
     if len(template.messages) % 2 != 0:
@@ -245,10 +243,8 @@ def tokenize_rlhf(
 
     for idx, mess in enumerate(context):
         if mess["from"] != template.roles[idx % 2]:
-            raise ValueError(
-                f"Message should iterate between user and assistant and starts with a \
-                             line from the user. Got the following data:\n{context}"
-            )
+            raise ValueError(f"Message should iterate between user and assistant and starts with a \
+                             line from the user. Got the following data:\n{context}")
         template.append_message(mess["from"], mess["content"])
 
     if len(template.messages) % 2 != 1:
@@ -272,18 +268,14 @@ def tokenize_rlhf(
     rejected_continuation = data_point["rejected"]
     for round in range(len(chosen_continuation)):
         if chosen_continuation[round]["from"] != template.roles[(round + 1) % 2]:
-            raise ValueError(
-                f"Message should iterate between user and assistant and starts with a \
-                             line from the user. Got the following data:\n{chosen_continuation}"
-            )
+            raise ValueError(f"Message should iterate between user and assistant and starts with a \
+                             line from the user. Got the following data:\n{chosen_continuation}")
         chosen.append_message(chosen_continuation[round]["from"], chosen_continuation[round]["content"])
 
     for round in range(len(rejected_continuation)):
         if rejected_continuation[round]["from"] != template.roles[(round + 1) % 2]:
-            raise ValueError(
-                f"Message should iterate between user and assistant and starts with a \
-                             line from the user. Got the following data:\n{rejected_continuation}"
-            )
+            raise ValueError(f"Message should iterate between user and assistant and starts with a \
+                             line from the user. Got the following data:\n{rejected_continuation}")
         rejected.append_message(rejected_continuation[round]["from"], rejected_continuation[round]["content"])
 
     (
@@ -296,14 +288,14 @@ def tokenize_rlhf(
     ) = (None, None, None, None, None, None)
 
     chosen_data_packed = apply_rlhf_data_format(chosen, tokenizer)
-    (chosen_input_ids, chosen_loss_mask, chosen_label_decode) = (
+    chosen_input_ids, chosen_loss_mask, chosen_label_decode = (
         chosen_data_packed["input_ids"],
         chosen_data_packed["loss_mask"],
         chosen_data_packed["label_decode"],
     )
 
     rejected_data_packed = apply_rlhf_data_format(rejected, tokenizer)
-    (rejected_input_ids, rejected_loss_mask, rejected_label_decode) = (
+    rejected_input_ids, rejected_loss_mask, rejected_label_decode = (
         rejected_data_packed["input_ids"],
         rejected_data_packed["loss_mask"],
         rejected_data_packed["label_decode"],

applications/ColossalChat/coati/distributed/reward/reward_fn.py

@@ -17,7 +17,6 @@
 https://github.com/volcengine/verl
 """
 
-
 import json
 
 import torch

applications/ColossalChat/coati/trainer/kto.py

@@ -130,7 +130,7 @@ def _train(self, epoch: int):
         )
         for i, batch in enumerate(self.train_dataloader):
             batch = to_device(batch, self.device)
-            (input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask) = (
+            input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask = (
                 batch["input_ids"],
                 batch["attention_mask"],
                 batch["loss_mask"],
@@ -279,7 +279,7 @@ def _eval(self, epoch: int):
         )
         for i, batch in enumerate(self.train_dataloader):
             batch = to_device(batch, self.device)
-            (input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask) = (
+            input_ids, attention_mask, loss_mask, label, kl_input_ids, kl_attention_mask, kl_loss_mask = (
                 batch["input_ids"],
                 batch["attention_mask"],
                 batch["loss_mask"],

applications/ColossalChat/examples/community/ray/train_prompts_on_ray.py

@@ -120,7 +120,7 @@ def _init_optimizer(self):
     def _prepare_model_with_strategy(self, has_optimizer: bool):
         if has_optimizer:
             self._init_optimizer()
-            (self._model, self._optimizer) = self._strategy.prepare((self._model, self._optimizer))
+            self._model, self._optimizer = self._strategy.prepare((self._model, self._optimizer))
         else:
             self._model = self._strategy.prepare(self._model)
 

applications/ColossalQA/examples/webui_demo/webui.py

@@ -81,11 +81,11 @@ def restart(chatbot, txt):
     )
     with gr.Row():
         btn = gr.UploadButton("📁", file_types=["file"], file_count="multiple", size="sm")
-        restart_btn = gr.Button(str("\u21BB"), elem_id="restart-btn", scale=1)
+        restart_btn = gr.Button(str("\u21bb"), elem_id="restart-btn", scale=1)
         txt = gr.Textbox(
             scale=8,
             show_label=False,
-            placeholder="Enter text and press enter, or use 📁 to upload files, click \u21BB to clear loaded files and restart chat",
+            placeholder="Enter text and press enter, or use 📁 to upload files, click \u21bb to clear loaded files and restart chat",
             container=True,
             autofocus=True,
         )

colossalai/auto_parallel/tensor_shard/solver/solver.py

@@ -1,6 +1,6 @@
 """This code is adapted from Alpa
-    https://github.com/alpa-projects/alpa/
-   with some changes. """
+ https://github.com/alpa-projects/alpa/
+with some changes."""
 
 import multiprocessing
 import time

colossalai/autochunk/select_chunk.py

@@ -176,7 +176,7 @@ def _select_min_memory_chunk_region(self, possible_chunk_regions, chunk_infos):
         return best_region
 
     def _is_legal_region(self, cur_chunk_info, chunk_infos):
-        (chunk_region_start, chunk_region_end) = cur_chunk_info["region"]
+        chunk_region_start, chunk_region_end = cur_chunk_info["region"]
         if cur_chunk_info in chunk_infos:
             return False
         if chunk_region_end < chunk_region_start:

colossalai/booster/plugin/gemini_plugin.py

@@ -338,10 +338,8 @@ def load_sharded_optimizer(
         # Load param_groups.
         param_group_path = ckpt_index_file.get_param_group_filename()
         if param_group_path is None:
-            raise RuntimeError(
-                f"Invalid index file path {checkpoint_index_file} for an optimizer. \
-                               Lacking param group file under current directory."
-            )
+            raise RuntimeError(f"Invalid index file path {checkpoint_index_file} for an optimizer. \
+                               Lacking param group file under current directory.")
         saved_param_groups = torch.load(param_group_path)
         optimizer.load_param_groups(saved_param_groups)
 

colossalai/booster/plugin/low_level_zero_plugin.py

@@ -268,10 +268,8 @@ def load_sharded_optimizer(
         # Load param_groups
         param_group_path = ckpt_index_file.get_param_group_filename()
         if param_group_path is None:
-            raise RuntimeError(
-                f"Invalid index file path {index_file_path} for an optimizer. \
-                               Lacking param group file under current directory."
-            )
+            raise RuntimeError(f"Invalid index file path {index_file_path} for an optimizer. \
+                               Lacking param group file under current directory.")
         id_map = load_param_groups_into_optimizer(optimizer, param_group_path)
 
         checkpoint_files, _ = ckpt_index_file.get_checkpoint_filenames()