Hi @meganomaly, I would like to fine-tune the following model for the competition : Model: https://huggingface.co/Kansallisarkisto/multicentury-htr-model - TrO... See more

We're sitting around 0.906 and have spent the last stretch testing ideas that soundedobvious and turned out to be worth nothing. Posting the negative results so... See more

Thanks to everyone who has raised questions around external data and pretrained models. For this challenge, we want to keep the final solutions reproducible and... See more

Following the ruling on stanford-oval/churro-3B (denied because the base model carried a non-commercial Qwen Research License), I would like to ask for a clarif... See more


I wanted clarification on whether it is allowed to use other external datasets that are publicly available and mainted such as MNIST,EMNIST,READ / ICFHR dataset... See more

Hi — a rules clarification, following the helpful answers in #33870, #33891 and #34459.Concretely:1. Segment character-level crops from the provided Train.csv i... See more

Hi @meganomaly, On 5 Aug you said models under research-only or non-commercial licences need reviewing before use, and asked entrants to post the model name an... See more

Hello, I have a quick question about the "pretrained models must be openly available to everyone" rule.On behalf of my team on this very interesting task, I wr... See more

Transductive pseudo-labeling / self-training on the test images is allowed, provided the entire process is fully automated. To ensure fairness and transparency... See more

Hi!Do the open-source-tools and no-proprietary-model-service rules prohibit using OpenAI Codex without sharing competition data? If not, what disclosure is requ... See more

Hi everyone! I have a question for folks who broke above .91, I am not looking for exact setups, just curious about the general direction people went. Are you m... See more

hi

@meganomaly Does the image-metadata rule prohibit standard aspect-ratio-preserving resize and padding when dimensions are not supplied as predictive features?

I have CER 75% but reeaching 98+ by external datasets of relevant same hadwritten English. It is open public datasets. I see it is forbidden so will submit pure... See more

Hi Zindi team, I am reopening this clarification because the earlier question, "Rules clarification: self-training / pseudo-labeling on the test images," remai... See more

Hi there,Just joined this competiton.There seems to be an issue downloading the image.zip file. It downloads 23.2 MB then freezes. Can this be investigated? Or... See more

Hi Zindi team,A rules clarification on using the test set during training.Is it permitted to use the provided test images in training via self-training / transd... See more

Hi all,While going through the training images, I noticed that most multi-line images are labeled with only the text of the centered/main line, but a portion of... See more

The rules allow pretrained models that are openly available to everyone. Does that include models released under research/non-commercial licenses (for example t... See more

If you look in the code of their Starter code, you will see a file: eval_metrics.py which have an import at the topimport osimport sysimport yamlimport pandas a... See more

Hello everyone, I hope you're doing well. I have two quick questions about the rules for the ROAD Barbados Historic Handwriting Challenge: Are hyperparameter op... See more

Are participants permitted to rent raw GPU compute from RunPod, AWS, or similar solely to train and infer publicly available open-source models using only the p... See more

Can you share the exact wer.py/cer.py implementation used for scoring, or clarify why WER Weighted values exceed 1.0? Trying to align local validation with lead... See more

Barbados’ colonial-era history is preserved in thousands of handwritten pages - deeds, wills, estate inventories, census records, and other archival documents d... See more