发布节奏
最近 12 周每周中位浏览
最近 12 周发布时间
颜色 = 中位浏览爆款 vs 日常
爆款 = 浏览前 10%| 指标 | 爆款 | 日常 |
|---|---|---|
| 帖子 | 4 | 29 |
| 带媒体 | 100% | 41.4% |
| 回复 | 0% | 0% |
| 平均字数 | 271 | 197 |
| 中位浏览 | 5,097 | 291 |
打法建议
只陈述相关性- 互动最好的时段是 15:00 前后（北京时间），中位浏览 2,061。
- 一周里 周一 的中位浏览最高。
- 形式上 纯文字 的中位浏览更高（300 vs 379）。
- 爆款里带媒体占 100%、回复占 0%；日常分别是 41.4%、0%。
- 当前节奏约 0.7 条/天，日均浏览 1,197。
内容结构
按本站分析的分类（近 120 天）还没有分类分析。
形式与语言
帖子占比- 纯文字17
- 图片15
- 视频1
回复占比 0% · 引用 57.6% · 带媒体 48.5%
互动最高的帖子
按浏览排序- 1
🤗 @GoogleDeepMind is back with EmbeddingGemma 2, which maps text (including code), images, video & audio into one shared 768-dimensional embedding space. 740M total parameters, 100+ languages, Apache 2.0 & Sentence Transformers support. Thread with highlights 🧵
- 2
Not as sexy as generative decoders, but Sentence Transformers models for search have slowly taken over the most downloaded @huggingface models page over the last few years. I'm hoping to see a further resurgence of encoders with the new interest in Jevlike zero-shot encoders.
- 3
The original author of the SPLADE line (THE sparse embedding architecture imo), just released the biggest advancement of Sparse embedding models since SPLADE-v3! - 150M, beats every sparse model up to 1B at search - Reaches sub-ms latency using Seismic (!) Go @Linkup_platform !
- 4
🔎 @perplexity_ai just released pplx-embed-v2-late: 0.6B & 9B models for text & image retrieval, natively supported in Sentence Transformers. Search document pages like PDFs, slide decks, screenshots, charts, etc. as images, without OCR. Thread with highlights 🧵
- 5
I'm so glad that Google Deepmind chose an accessible model size for their excellent multimodal EmbeddingGemma 2! Plus some clever work to make audio and video faster to process. All nicely integrated in Sentence Transformers as well 👏
- 6
Oooh, I think this is a great option. EG2 is very small, efficient, and multimodal. It's also quite efficient with videos, not processing every single frame.
- 7
And Linkup's SPARSEUP beats existing SparseEncoder models by quite a huge gap at general-purpose retrieval/search as well, impressive!
- 8
It's honestly a lot of fun to finetune
- 9
My all-time favourite demo for multimodal embedding search: enter "Birds" and find images, videos, and even audio snippets of birds! Or search with images or sound instead. https://huggingface.co/spaces/webml-community/embeddinggemma-2-webgpu Featuring yesterday's @GoogleDeepMind EmbeddingGemma 2 and WebGPU
- 10
📈@Google's EmbeddingGemma2 is looking to be a huge success! After only a week, it has surpasses EG1's likes to become the 4th most liked Google model ever, as well as the 3rd most liked Sentence Transformers model ever. Congratulations to the team for their work! Links in 🧵
- 11
Here's the open decision model links in case you don't want to scroll all the way down this thread: - https://huggingface.co/LiquidAI/d1-3B - https://huggingface.co/LiquidAI/d1-omni-600M Thank me later.
- 12
I always love models trained for one non-English models, and it's no exception here. @BdsLoick from @AI_AlphaEdge trained a French Sparse Embedding model for retrieval. It's SOTA at its size (150M), very impressive as sparse embedding models are notoriously hard to train. ->🧵
点赞、转发、浏览是采集那一刻的快照，不是帖子最终数据；时间为北京时间。KPI 卡片里的同比只在账号有 8 周以上数据时显示， 否则只看近 4 周。“打法建议”只说明数据里的相关性，不代表因果。