apunkt/MetaGPT
1
0
Fork 0
mirror of https://github.com/FoundationAgents/MetaGPT.git synced 2026-06-05 14:55:18 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

change img path from abs to rel

Browse source
This commit is contained in:
Yizhou Chi 2024-09-13 19:04:48 +08:00
parent a6b066a127
commit cfa21ba27e
1 changed files with 28 additions and 28 deletions
Download patch file Download diff file Expand all files Collapse all files

56
expo/data/hf_data.py
View file

@ -1,19 +1,19 @@
import asyncio
import io
import os
from pathlib import Path
import numpy as np
from PIL import Image
import io
import pandas as pd
from datasets import load_dataset
from PIL import Image
from expo.data.dataset import ExpDataset, process_dataset, save_datasets_dict_to_yaml
from expo.insights.solution_designer import SolutionDesigner
HFDATSETS = [
# {"name": "sms_spam", "dataset_name": "ucirvine/sms_spam", "target_col": "label", "modality": "text"},
# {"name": "banking77", "dataset_name": "PolyAI/banking77", "target_col": "label", "modality": "text"},
# {"name": "gnad10", "dataset_name": "community-datasets/gnad10", "target_col": "label", "modality": "text"},
{"name": "sms_spam", "dataset_name": "ucirvine/sms_spam", "target_col": "label", "modality": "text"},
{"name": "banking77", "dataset_name": "PolyAI/banking77", "target_col": "label", "modality": "text"},
{"name": "gnad10", "dataset_name": "community-datasets/gnad10", "target_col": "label", "modality": "text"},
{
"name": "oxford-iiit-pet",
"dataset_name": "timm/oxford-iiit-pet",
@ -21,18 +21,20 @@ HFDATSETS = [
"target_col": "label_cat_dog",
"modality": "image",
},
# { "name": "stanford_cars",
# "dataset_name": "tanganke/stanford_cars",
# "image_col": "image",
# "target_col": "label",
# "modality": "image"},
# {
# "name": "fashion_mnist",
# "dataset_name": "zalando-datasets/fashion_mnist",
# "image_col": "image",
# "target_col": "label",
# "modality": "image",
# },
{
"name": "stanford_cars",
"dataset_name": "tanganke/stanford_cars",
"image_col": "image",
"target_col": "label",
"modality": "image",
},
{
"name": "fashion_mnist",
"dataset_name": "zalando-datasets/fashion_mnist",
"image_col": "image",
"target_col": "label",
"modality": "image",
},
]
@ -81,17 +83,19 @@ class HFExpDataset(ExpDataset):
return df, test_df
def save_images_and_update_df(self, df, raw_dir, split):
image_dir = Path(raw_dir, f"{split}_images")
image_dir.mkdir(parents=True, exist_ok=True)
abs_image_dir = Path(raw_dir, f"{split}_images")
rel_image_dir = f"raw/{split}_images"
abs_image_dir.mkdir(parents=True, exist_ok=True)
def process_image(idx, row):
image_bytes = row[self.image_col]["bytes"]
image = Image.open(io.BytesIO(image_bytes))
if image.mode == "RGBA":
image = image.convert("RGB")
img_path = Path(image_dir, f"{idx}.jpg")
img_path = Path(abs_image_dir, f"{idx}.jpg")
rel_img_path = f"{rel_image_dir}/{idx}.jpg"
image.save(img_path)
return str(img_path)
return rel_img_path
df["image"] = df.apply(lambda row: process_image(row.name, row), axis=1)
return df
@ -112,7 +116,7 @@ class HFExpDataset(ExpDataset):
if __name__ == "__main__":
dataset_dir = "D:/work/automl/datasets"
save_analysis_pool = True
force_update = False
force_update = True
datasets_dict = {"datasets": {}}
solution_designer = SolutionDesigner()
for dataset_meta in HFDATSETS:
@ -125,9 +129,5 @@ if __name__ == "__main__":
force_update=force_update,
modality=dataset_meta["modality"],
)
asyncio.run(
process_dataset(
hf_dataset, solution_designer, save_analysis_pool, datasets_dict
)
)
asyncio.run(process_dataset(hf_dataset, solution_designer, save_analysis_pool, datasets_dict))
save_datasets_dict_to_yaml(datasets_dict, "hf_datasets.yaml")