We describe LiLa, our submission to the constrained track of the WMT 2026 General Machine Translation Task. Motivated by recent advances in large language models (LLMs) and their increasing availability as open models, we investigate the adaptation of LLMs for Ligurian and Ladin, two low-resource languages included in the shared task. We evaluate several adaptation strategies for Gemma-4, including Instruction Tuning (IT-only), continuous pre-training followed by Instruction Tuning (CPT+IT), and Low-Rank Adaptation (LoRA), and compare them with a dedicated encoder-decoder pipeline based on fine-tuned No-Language Left Behind (NLLB) models. Our experiments show that, under the evaluated conditions, fine-tuned NLLB models achieve higher automatic translation scores than the investigated LLM adaptation strategies. Based on these results, our final submission adopts a fine-tuned NLLB-based translation pipeline. We discuss the main observations from the different approaches explored during development. We release the code and models produced in these experiments through a publicly accessible GitHub repository https://github.com/schtailmuel/wmt26-lila/

Culturally grounded image captioning for under-resourced Indigenous languages is challenging due to severe data scarcity and the need to describe culturally specific visual content. This paper describes our submission to the AmericasNLP 2026 shared task, where we evaluate two architectural paradigms for caption generation across Bribri, Guaraní, Yucatec Maya, Wixárika, and Orizaba Nahuatl. First, we implement a cascaded system that combines a large vision-language model with a machine translation pipeline, showing that culturally contextualized, persona-based prompting improves over the official baseline in most comparable settings. Second, we develop a direct, end-to-end Single-stage approach by adapting PaliGemma 2 using LoRA fine-tuning, continued pre-training, and multilingual joint training. Our single-stage experiments show that, despite severe domain mismatch and reliance on synthetic training data, multilingual training and continued pre-training improve automatic chrF++ relative to single-language LoRA fine-tuning in some settings. Overall, cascaded pipelines remain the strongest among the evaluated approaches under current data constraints, while single-stage models remain a promising but currently data-limited path toward direct Indigenous-language image captioning.