SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.04007 · 2024

SEED-Data-Edit Technical Report: A Hybrid Dataset for Instructional Image Editing

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 9 functions out of this paper's own repositories and ran 7 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
ailab-cvc/seed-x canonical 7 of 9
FunctionStatusWhere it lives
cosine_loss Ran ailab-cvc/seed-x/src/models/mllm/seed_x.py
pointer only (licence: NONE) · get_code("66a2fe5698649c4c")
extract_box Ran ailab-cvc/seed-x/src/inference/eval_img2text_seed_x.py
pointer only (licence: NOASSERTION) · get_code("c3e1f507b8bc217d")
get_transform Ran ailab-cvc/seed-x/src/processer/transforms.py
pointer only (licence: NOASSERTION) · get_code("bda035480e718551")
resize_and_pad_image Ran ailab-cvc/seed-x/src/inference/any_res.py
pointer only (licence: NOASSERTION) · get_code("4039ff7e52e2d4fb")
select_best_resolution Ran ailab-cvc/seed-x/src/inference/any_res.py
pointer only (licence: NONE) · get_code("3999ff487573f32c")
select_best_resolution_v2 Ran ailab-cvc/seed-x/src/inference/any_res.py
pointer only (licence: NOASSERTION) · get_code("4db1fefd729f4ef0")
trainable_params Ran ailab-cvc/seed-x/src/train/train_seed_x_sft.py
pointer only (licence: NOASSERTION) · get_code("070e1db10f514798")
bert_tokenizer Not yet run ailab-cvc/seed-x/src/processer/tokenizer.py
pointer only (licence: NOASSERTION) · get_code("04c1966e1890614b")
extract_box Not yet run ailab-cvc/seed-x/src/inference/eval_img2text_seed_x_i.py
pointer only (licence: NOASSERTION) · get_code("685a977b9afc9b0a")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In this technical report, we introduce SEED-Data-Edit: a unique hybrid dataset for instruction-guided image editing, which aims to facilitate image manipulation using open-form language. SEED-Data-Edit is composed of three distinct types of data: (1) High-quality editing data produced by an automated pipeline, ensuring a substantial volume of diverse image editing pairs. (2) Real-world scenario data collected from the internet, which captures the intricacies of user intentions for promoting the practical application of image editing in the real world. (3) High-precision multi-turn editing data annotated by humans, which involves multiple rounds of edits for simulating iterative editing processes. The combination of these diverse data sources makes SEED-Data-Edit a comprehensive and versatile dataset for training language-guided image editing model. We fine-tune a pretrained Multimodal Large Language Model (MLLM) that unifies comprehension and generation with SEED-Data-Edit. The instruction tuned model demonstrates promising results, indicating the potential and effectiveness of SEED-Data-Edit in advancing the field of instructional image editing. The datasets are released in https://huggingface.co/datasets/AILab-CVC/SEED-Data-Edit.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2405.04007")
get_code_for_paper("2405.04007")
have("2405.04007")

Connect an agent — have() is free.