make format happy

aria/data.py

@@ -22,9 +22,10 @@
 from typing import Dict, Iterable, List
 
 import torch
-from datasets import DatasetDict, concatenate_datasets, load_dataset
 from datasets.features import Features, Sequence, Value
 
+from datasets import DatasetDict, concatenate_datasets, load_dataset
+
 
 def apply_chat_template_and_tokenize(
     messages_batch: List[List[Dict]],
@@ -96,7 +97,7 @@ def create_target(role, input_id):
         if pad_length > 0:
             input_ids[i] = input_ids[i] + [tokenizer.pad_token_id] * pad_length
             targets[i] = targets[i] + [IGNORE_TOKEN_ID] * pad_length
-        else: # truncate
+        else:  # truncate
             input_ids[i] = input_ids[i][:max_batch_len]
             targets[i] = targets[i][:max_batch_len]
 

aria/train.py

@@ -114,7 +114,13 @@ def setup_peft(model, model_config):
     return model
 
 
-def collate_fn(examples, tokenizer, processor, split_image: bool = False, max_seq_length: int = 1024):
+def collate_fn(
+    examples,
+    tokenizer,
+    processor,
+    split_image: bool = False,
+    max_seq_length: int = 1024,
+):
     images = []
     messages = []
     for example in examples:
@@ -172,7 +178,7 @@ def collate_fn(examples, tokenizer, processor, split_image: bool = False, max_se
                             message["content"].insert(cont_idx + img_i, insert_item)
             messages.append(example["messages"])
         else:
-            if example['images']:
+            if example["images"]:
                 images.extend(example["images"])
             messages.append(example["messages"])
 
@@ -192,7 +198,7 @@ def collate_fn(examples, tokenizer, processor, split_image: bool = False, max_se
 
         batch.update(image_inputs)
         batch["pixel_values"] = batch["pixel_values"].to(torch.bfloat16)
-    else: # text-only
+    else:  # text-only
         batch = apply_chat_template_and_tokenize(
             messages,
             tokenizer,
@@ -225,7 +231,11 @@ def main():
         model=model,
         args=training_args,
         data_collator=lambda examples: collate_fn(
-            examples, tokenizer, processor, model_config.split_image, training_args.max_seq_length
+            examples,
+            tokenizer,
+            processor,
+            model_config.split_image,
+            training_args.max_seq_length,
         ),
         train_dataset=train_dataset,
         eval_dataset=eval_dataset,

examples/code_sft/evaluation.py

@@ -1,11 +1,12 @@
-from human_eval.evaluation import evaluate_functional_correctness
-import json
 import argparse
-
 import os
 
+from human_eval.evaluation import evaluate_functional_correctness
+
 parser = argparse.ArgumentParser()
-parser.add_argument("--save_root", required=True, type=str, help="the result directory of humaneval")
+parser.add_argument(
+    "--save_root", required=True, type=str, help="the result directory of humaneval"
+)
 args = parser.parse_args()
 
 tmp_dir = os.path.join(args.save_root, "tmp")
@@ -20,4 +21,4 @@
     language="python",
 )
 
-print(result)
+print(result)

examples/code_sft/get_data.py

@@ -1,28 +1,33 @@
-import os
 import json
+import os
+
 from tqdm import trange
 
 from datasets import load_dataset
 
-save_dir = './datasets/code_sft'
+save_dir = "./datasets/code_sft"
 os.makedirs(save_dir, exist_ok=True)
 
-dataset = load_dataset("ise-uiuc/Magicoder-Evol-Instruct-110K")['train']
+dataset = load_dataset("ise-uiuc/Magicoder-Evol-Instruct-110K")["train"]
 l_ds = len(dataset)
 
-with open(os.path.join(save_dir, "train.jsonl"),"w") as fo:
+with open(os.path.join(save_dir, "train.jsonl"), "w") as fo:
     for i in trange(l_ds):
-        instruction, response = dataset[i]['instruction'], dataset[i]['response']
+        instruction, response = dataset[i]["instruction"], dataset[i]["response"]
         item = {
             "messages": [
-                    {
-                        "content": [{"text": instruction, "type": "text"},],
-                        "role": "user"
-                    },
-                    {
-                        "content": [{"text": response, "type": "text"},],
-                        "role": "assistant"
-                    },
-                ],
+                {
+                    "content": [
+                        {"text": instruction, "type": "text"},
+                    ],
+                    "role": "user",
+                },
+                {
+                    "content": [
+                        {"text": response, "type": "text"},
+                    ],
+                    "role": "assistant",
+                },
+            ],
         }
         fo.write(f"{json.dumps(item, ensure_ascii=False)}\n")

examples/code_sft/human_eval/data.py

@@ -1,8 +1,7 @@
-from typing import Iterable, Dict
 import gzip
 import json
 import os
-
+from typing import Dict, Iterable
 
 ROOT = os.path.dirname(os.path.abspath(__file__))
 HUMAN_EVAL = os.path.join(ROOT, "..", "data", "HumanEval.jsonl.gz")
@@ -18,7 +17,7 @@ def stream_jsonl(filename: str) -> Iterable[Dict]:
     """
     if filename.endswith(".gz"):
         with open(filename, "rb") as gzfp:
-            with gzip.open(gzfp, 'rt') as fp:
+            with gzip.open(gzfp, "rt") as fp:
                 for line in fp:
                     if any(not x.isspace() for x in line):
                         yield json.loads(line)
@@ -34,16 +33,16 @@ def write_jsonl(filename: str, data: Iterable[Dict], append: bool = False):
     Writes an iterable of dictionaries to jsonl
     """
     if append:
-        mode = 'ab'
+        mode = "ab"
     else:
-        mode = 'wb'
+        mode = "wb"
     filename = os.path.expanduser(filename)
     if filename.endswith(".gz"):
         with open(filename, mode) as fp:
-            with gzip.GzipFile(fileobj=fp, mode='wb') as gzfp:
+            with gzip.GzipFile(fileobj=fp, mode="wb") as gzfp:
                 for x in data:
-                    gzfp.write((json.dumps(x) + "\n").encode('utf-8'))
+                    gzfp.write((json.dumps(x) + "\n").encode("utf-8"))
     else:
         with open(filename, mode) as fp:
             for x in data:
-                fp.write((json.dumps(x) + "\n").encode('utf-8'))
+                fp.write((json.dumps(x) + "\n").encode("utf-8"))

examples/code_sft/human_eval/evaluate_functional_correctness.py

@@ -1,7 +1,7 @@
-import fire
 import sys
 
-from .data import HUMAN_EVAL
+import fire
+
 from .evaluation import evaluate_functional_correctness
 
 
@@ -18,7 +18,9 @@ def entry_point(
     results to f"{sample_file}_results.jsonl.gz"
     """
     k = list(map(int, k.split(",")))
-    results = evaluate_functional_correctness(sample_file, k, n_workers, timeout, problem_file, is_mbpp)
+    results = evaluate_functional_correctness(
+        sample_file, k, n_workers, timeout, problem_file, is_mbpp
+    )
     print(results)