# Latest **URL:** https://community.cloud.ai4eosc.eu/latest.md [Latest](https://community.cloud.ai4eosc.eu/latest.md) · [Categories](https://community.cloud.ai4eosc.eu/categories.md) · [Tags](https://community.cloud.ai4eosc.eu/tags.md) --- ## [Welcome to AI community in EOSC! 👋](https://community.cloud.ai4eosc.eu/t/welcome-to-ai-community-in-eosc/5)
**Author:** [@system](https://community.cloud.ai4eosc.eu/u/system)\ **Replies:** 0\ **Last updated:** [June 3, 2025, 10:06pm UTC](https://community.cloud.ai4eosc.eu/t/welcome-to-ai-community-in-eosc/5 "2025-06-03T22:06:00Z")
We are so glad you joined us. AI community in EOSC. Here are some things you can do to get started: :speaking\_head: Introduce yourself by adding your picture and information about yourself and your interests to your… --- ## [LLM-as-a-Judge: Does Your Judge Need to Be Stronger Than the Model It Evaluates?](https://community.cloud.ai4eosc.eu/t/llm-as-a-judge-does-your-judge-need-to-be-stronger-than-the-model-it-evaluates/69)
**Author:** [@Fahimeh\_Alibabaei](https://community.cloud.ai4eosc.eu/u/Fahimeh_Alibabaei)\ **Replies:** 0\ **Last updated:** [November 3, 2026, 5:50pm UTC](https://community.cloud.ai4eosc.eu/t/llm-as-a-judge-does-your-judge-need-to-be-stronger-than-the-model-it-evaluates/69 "2026-11-03T17:50:24Z")
Evaluating LLM outputs at scale is one of the most practical problems in GenAI work. Human annotation is slow and expensive, and classic metrics such as BLEU or ROUGE say little about whether an answer is actually correc… --- ## [Reproducing BigEarthNet.txt and taking it federated: a multimodal vision-language model trained with NVFlare](https://community.cloud.ai4eosc.eu/t/reproducing-bigearthnet-txt-and-taking-it-federated-a-multimodal-vision-language-model-trained-with-nvflare/68)
**Author:** [@Leo](https://community.cloud.ai4eosc.eu/u/Leo)\ **Replies:** 0\ **Last updated:** [November 3, 2026, 3:21pm UTC](https://community.cloud.ai4eosc.eu/t/reproducing-bigearthnet-txt-and-taking-it-federated-a-multimodal-vision-language-model-trained-with-nvflare/68 "2026-11-03T15:21:32Z")
TL;DR: We re-implemented and reproduced the RS-InternVL baseline from the BigEarthNet.txt paper and exceeded the reported results on three of four tasks. We then trained the same multimodal model federated with NVIDIA FL… --- ## [Choices for inference](https://community.cloud.ai4eosc.eu/t/choices-for-inference/44)
**Author:** [@Peter\_Sisan](https://community.cloud.ai4eosc.eu/u/Peter_Sisan)\ **Replies:** 1\ **Last updated:** [August 18, 2025, 10:16am UTC](https://community.cloud.ai4eosc.eu/t/choices-for-inference/44 "2025-08-18T10:16:53Z")
Hi. I have my pre-trained model in a Docker container and I want to run it for inference. What choices do I have? --- ## [Reproducibility and provenance in AI workflows](https://community.cloud.ai4eosc.eu/t/reproducibility-and-provenance-in-ai-workflows/39)
**Author:** [@Irina\_Malkin\_Ondik](https://community.cloud.ai4eosc.eu/u/Irina_Malkin_Ondik)\ **Replies:** 2\ **Last updated:** [August 18, 2025, 9:17am UTC](https://community.cloud.ai4eosc.eu/t/reproducibility-and-provenance-in-ai-workflows/39 "2025-08-18T09:17:13Z")
Hi. How do you manage reproducibility and provenance in AI workflows? Are there tools that track both code, data, and model lineage, ideally in an open science-friendly way? --- ## [Good practices for publishing machine learning models and datasets in a FAIR-compliant manner](https://community.cloud.ai4eosc.eu/t/good-practices-for-publishing-machine-learning-models-and-datasets-in-a-fair-compliant-manner/38)
**Author:** [@Irina\_Malkin\_Ondik](https://community.cloud.ai4eosc.eu/u/Irina_Malkin_Ondik)\ **Replies:** 2\ **Last updated:** [August 7, 2025, 12:29pm UTC](https://community.cloud.ai4eosc.eu/t/good-practices-for-publishing-machine-learning-models-and-datasets-in-a-fair-compliant-manner/38 "2025-08-07T12:29:33Z")
Hi. What are good practices for publishing machine learning models and datasets in a FAIR-compliant manner? We are looking for guidance or examples, particularly in the context of EU research projects. --- ## [Publishing training datasets](https://community.cloud.ai4eosc.eu/t/publishing-training-datasets/48)
**Author:** [@evollmer](https://community.cloud.ai4eosc.eu/u/evollmer)\ **Replies:** 1\ **Last updated:** [August 7, 2025, 12:28pm UTC](https://community.cloud.ai4eosc.eu/t/publishing-training-datasets/48 "2025-08-07T12:28:35Z")
Hi! I want to publish my developed AI model training dataset and was wondering what options are available? --- ## [Distributed training without data sharing](https://community.cloud.ai4eosc.eu/t/distributed-training-without-data-sharing/47)
**Author:** [@evollmer](https://community.cloud.ai4eosc.eu/u/evollmer)\ **Replies:** 1\ **Last updated:** [August 6, 2025, 2:29pm UTC](https://community.cloud.ai4eosc.eu/t/distributed-training-without-data-sharing/47 "2025-08-06T14:29:42Z")
Hi! We are working on an application where our model training could benefit from some of our partner’s datasets, but they can’t share them with us due to privacy concerns. Are there AI methods and frameworks to execute d… --- ## [AI model improvement](https://community.cloud.ai4eosc.eu/t/ai-model-improvement/43)
**Author:** [@Peter\_Sisan](https://community.cloud.ai4eosc.eu/u/Peter_Sisan)\ **Replies:** 0\ **Last updated:** [August 5, 2025, 8:47pm UTC](https://community.cloud.ai4eosc.eu/t/ai-model-improvement/43 "2025-08-05T20:47:27Z")
I want to improve an existing AI model and need more resources to improve it, where can I find them? --- ## [Best practices labelling/benchmarking of RAG chatbot systems?](https://community.cloud.ai4eosc.eu/t/best-practices-labelling-benchmarking-of-rag-chatbot-systems/37)
**Author:** [@Irina\_Malkin\_Ondik](https://community.cloud.ai4eosc.eu/u/Irina_Malkin_Ondik)\ **Replies:** 0\ **Last updated:** [August 5, 2025, 11:37am UTC](https://community.cloud.ai4eosc.eu/t/best-practices-labelling-benchmarking-of-rag-chatbot-systems/37 "2025-08-05T11:37:51Z")
Hi. Are there any best practices for development and especially labelling/benchmarking of RAG chatbot systems? We have found out, that labelling of every new version/architecture of chatbot by human is quite demanding a…