Unverified50% confidenceFactExact time
Sona's offline tokenizer starts from a frozen Qwen2.5-Omni multimodal model and takes two types of input: the mel spectrogram of the first 90 seconds of a track and the track's metadata.
1
Sources
50%
Confidence
Long-term
Relevance
10/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedQwen3.5-Omni采用原生多模态预训练方式，从数据构建阶段就将图像、音频、视频与文本混合编排，与早期'感知模块+语言模型'串联方案有本质区别60% similarUnverifiedThe core idea behind Sona is to have the decoder directly generate Semantic IDs assigned to each track, rather than performing traditional similarity-based retrieval over a massive music library.58% similarUnverifiedSecond-generation sequencing is represented by Illumina's sequencing-by-synthesis technology, processing short-read fragments typically 150–300 bp in length with an accuracy rate above 99.9%.57% similarUnverifiedQwen is a series of large language models independently developed by Alibaba Cloud, covering versions ranging from 1.8B to 110B parameters and supporting tasks such as text generation, code writing, and multimodal understanding.56% similarUnverified对 PaliGemma 2 检测任务微调通常建议冻结视觉编码器权重、仅微调语言模型部分55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/980326API
curl https://kongchang.com/api/v1/knowledge/claims/980326MCP
get_claim(id=980326)