Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.22943 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914417319346176 |
|---|---|
| author | Wang, Qirui Guo, Qi Sun, Yiding Yang, Junkai Zhang, Dongxu Pang, Shanmin Guo, Qing |
| author_facet | Wang, Qirui Guo, Qi Sun, Yiding Yang, Junkai Zhang, Dongxu Pang, Shanmin Guo, Qing |
| contents | Personalized text-to-image generation lets users fine-tune diffusion models into repositories of concept-specific checkpoints, but serving these repositories efficiently is difficult for two reasons: natural-language requests are often ambiguous and can be misrouted to visually similar checkpoints, and standard post-training quantization can distort the fragile representations that encode personalized concepts. We present PersonalQ, a unified framework that connects checkpoint selection and quantization through a shared signal -- the checkpoint's trigger token. Check-in performs intent-aligned selection by combining intent-aware hybrid retrieval with LLM-based reranking over checkpoint context and asks a brief clarification question only when multiple intents remain plausible; it then rewrites the prompt by inserting the selected checkpoint's canonical trigger. Complementing this, Trigger-Aware Quantization (TAQ) applies trigger-aware mixed precision in cross-attention, preserving trigger-conditioned key/value rows (and their attention weights) while aggressively quantizing the remaining pathways for memory-efficient inference. Experiments show that PersonalQ improves intent alignment over retrieval and reranking baselines, while TAQ consistently offers a stronger compression-quality trade-off than prior diffusion PTQ methods, enabling scalable serving of personalized checkpoints without sacrificing fidelity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_22943 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PersonalQ: Select, Quantize, and Serve Personalized Diffusion Models for Efficient Inference Wang, Qirui Guo, Qi Sun, Yiding Yang, Junkai Zhang, Dongxu Pang, Shanmin Guo, Qing Artificial Intelligence Personalized text-to-image generation lets users fine-tune diffusion models into repositories of concept-specific checkpoints, but serving these repositories efficiently is difficult for two reasons: natural-language requests are often ambiguous and can be misrouted to visually similar checkpoints, and standard post-training quantization can distort the fragile representations that encode personalized concepts. We present PersonalQ, a unified framework that connects checkpoint selection and quantization through a shared signal -- the checkpoint's trigger token. Check-in performs intent-aligned selection by combining intent-aware hybrid retrieval with LLM-based reranking over checkpoint context and asks a brief clarification question only when multiple intents remain plausible; it then rewrites the prompt by inserting the selected checkpoint's canonical trigger. Complementing this, Trigger-Aware Quantization (TAQ) applies trigger-aware mixed precision in cross-attention, preserving trigger-conditioned key/value rows (and their attention weights) while aggressively quantizing the remaining pathways for memory-efficient inference. Experiments show that PersonalQ improves intent alignment over retrieval and reranking baselines, while TAQ consistently offers a stronger compression-quality trade-off than prior diffusion PTQ methods, enabling scalable serving of personalized checkpoints without sacrificing fidelity. |
| title | PersonalQ: Select, Quantize, and Serve Personalized Diffusion Models for Efficient Inference |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2603.22943 |