Spaces:
Runtime error
Runtime error
File size: 2,493 Bytes
8366b03 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 |
# Copyright 2023 Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import copy
import os
import json
from tqdm import tqdm
import ipdb
import random
from torch.nn.utils.rnn import pad_sequence
from dataclasses import dataclass, field
from typing import Callable, Dict, Sequence
import torch
import torch.distributed as dist
import transformers
from torch.utils.data import Dataset
from tqdm import tqdm
class SupervisedDataset(Dataset):
"""Dataset for supervised fine-tuning."""
def __init__(self, data_path: str, image_root_path: str):
super(SupervisedDataset, self).__init__()
with open(data_path, 'r') as f:
json_data = json.load(f)
# for debug:
#json_data = json_data[:100000]
self.image_path_list, self.caption_list = [], []
for item in json_data:
one_image_name, one_caption = item["image_name"], item["conversation"]
# TODO: stage 2 dataset format is invalid
if not one_image_name.endswith('.jpg'):
one_image_name += '.jpg'
one_image_path = image_root_path + '/{}'.format(one_image_name)
self.image_path_list.append(one_image_path)
self.caption_list.append(one_caption)
print(f'[!] collect {len(self.image_path_list)} samples for training')
def __len__(self): # number of instances
return len(self.image_path_list)
#def __getitem__(self, i) -> Dict[str, torch.Tensor]: # how to get item, 取一个样本
def __getitem__(self, i):
return dict(image_paths=self.image_path_list[i], output_texts=self.caption_list[i])
def collate(self, instances):
image_paths, output_texts = tuple([instance[key] for instance in instances] for key in ("image_paths", "output_texts"))
return dict(
image_paths=image_paths,
output_texts=output_texts
)
|