# Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved.
#
# 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 json
import random
from datasets import load_dataset
from nemo_automodel.datasets.vlm.utils import json2token
[docs]
def make_rdr_dataset(path_or_dataset="quintend/rdr-items", split="train", **kwargs):
"""Load and preprocess the RDR dataset for image-to-text fine-tuning.
Args:
path_or_dataset (str): Path or identifier for the RDR dataset.
split (str): Dataset split to load.
**kwargs: Additional arguments.
Returns:
Dataset: The processed dataset.
"""
dataset = load_dataset(path_or_dataset, split=split)
def format(example):
return {
"conversation": [
{
"role": "user",
"content": [
{"type": "image", "image": example["image"]},
{"type": "text", "text": "Describe this image."},
],
},
{
"role": "assistant",
"content": [{"type": "text", "text": example["text"]}],
},
],
}
return [format(example) for example in dataset]
# return dataset.map(format, batched=False)
[docs]
def make_cord_v2_dataset(
path_or_dataset="naver-clova-ix/cord-v2", split="train", **kwargs,
):
"""Load and preprocess the CORD-V2 dataset for image-to-text fine-tuning.
"""
dataset = load_dataset(path_or_dataset, split=split)
def format(example):
ground_truth = json.loads(example["ground_truth"])
if (
"gt_parses" in ground_truth
): # when multiple ground truths are available, e.g., docvqa
assert isinstance(ground_truth["gt_parses"], list)
gt_jsons = ground_truth["gt_parses"]
else:
assert "gt_parse" in ground_truth and isinstance(
ground_truth["gt_parse"], dict,
)
gt_jsons = [ground_truth["gt_parse"]]
text = random.choice(
[json2token(gt_json, sort_json_key=True) for gt_json in gt_jsons],
)
return {
"conversation": [
{
"role": "user",
"content": [
{"type": "image", "image": example["image"]},
{"type": "text", "text": "Describe this image."},
],
},
{"role": "assistant", "content": [{"type": "text", "text": text}]},
],
}
return [format(example) for example in dataset]
# return dataset.map(format, batched=False, num_proc=8,remove_columns=["ground_truth"])
