fast inference — 2 Technical Articles

Apr 22, 2025 · Artificial Intelligence

How DistilQwen2.5-DS3-0324 Achieves Fast, Accurate Reasoning via Quick‑Think Distillation

This article introduces DistilQwen2.5-DS3-0324, a distilled language model series that balances rapid inference with strong reasoning by applying a fast‑thinking chain‑of‑thought strategy, details its two‑stage distillation framework, evaluation on diverse benchmarks, and provides code for downloading and using the models.

chain of thoughtdeep learningfast inference

0 likes · 17 min read

How DistilQwen2.5-DS3-0324 Achieves Fast, Accurate Reasoning via Quick‑Think Distillation

Alibaba Cloud Big Data AI Platform

Jul 13, 2023 · Artificial Intelligence

Rapid Diffusion: Fast, Domain‑Specific Text‑to‑Image Generation for Chinese

Rapid Diffusion introduces a knowledge‑enhanced, high‑speed Chinese text‑to‑image diffusion model with one‑click deployment, achieving superior image quality and up to 1.73× faster inference through FlashAttention and BladeDISC optimizations, and demonstrates strong performance across e‑commerce, traditional painting, and food datasets.

Chinese NLPKnowledge Enhancementdiffusion model

0 likes · 12 min read

Rapid Diffusion: Fast, Domain‑Specific Text‑to‑Image Generation for Chinese