Spent the last few days working on another proof of concept for @standard.site topic-based discovery: longform.fyi
Possible categories are taken from the Ghost discovery feed, and categorization uses SentenceTransformers. I assume that eventually this will be done in-house by most AppViews
Possible categories are taken from the Ghost discovery feed, and categorization uses SentenceTransformers. I assume that eventually this will be done in-house by most AppViews
longform.fyi
Topical discovery for publications on the open social web
longform.fyi
February 19, 2026 at 8:02 PM
Spent the last few days working on another proof of concept for @standard.site topic-based discovery: longform.fyi
Possible categories are taken from the Ghost discovery feed, and categorization uses SentenceTransformers. I assume that eventually this will be done in-house by most AppViews
Possible categories are taken from the Ghost discovery feed, and categorization uses SentenceTransformers. I assume that eventually this will be done in-house by most AppViews
I know a lot of people are working on making ModernBERT-based embedding models, but in the meantime, if you’d like to play around with it (no better way to learn than practice), it’s plug&play with Sentence Transformers www.sbert.net and we have examples on the repo
SentenceTransformers Documentation — Sentence Transformers documentation
www.sbert.net
December 22, 2024 at 1:11 AM
I know a lot of people are working on making ModernBERT-based embedding models, but in the meantime, if you’d like to play around with it (no better way to learn than practice), it’s plug&play with Sentence Transformers www.sbert.net and we have examples on the repo
How are you with Python? You can run SentenceTransformers locally and it’s great for labeling and finetuning.
April 1, 2026 at 1:40 PM
How are you with Python? You can run SentenceTransformers locally and it’s great for labeling and finetuning.
Here check this out:
www.sbert.net
Try using similarity search with SQLite ^^
python.langchain.com/docs/integra...
www.sbert.net
Try using similarity search with SQLite ^^
python.langchain.com/docs/integra...
SentenceTransformers Documentation — Sentence Transformers documentation
www.sbert.net
July 11, 2025 at 5:12 AM
Here check this out:
www.sbert.net
Try using similarity search with SQLite ^^
python.langchain.com/docs/integra...
www.sbert.net
Try using similarity search with SQLite ^^
python.langchain.com/docs/integra...
Respectfully that, at best, is extremely dependent
Do you need Claude to reference ktp's HRT material and make it more accessible through translation, summarization, or rewording?
Absolutely not
The claudependent people throw stupid things like rewriting SentenceTransformers _but in rust_ as metrics
Do you need Claude to reference ktp's HRT material and make it more accessible through translation, summarization, or rewording?
Absolutely not
The claudependent people throw stupid things like rewriting SentenceTransformers _but in rust_ as metrics
August 20, 2026 at 7:49 PM
Respectfully that, at best, is extremely dependent
Do you need Claude to reference ktp's HRT material and make it more accessible through translation, summarization, or rewording?
Absolutely not
The claudependent people throw stupid things like rewriting SentenceTransformers _but in rust_ as metrics
Do you need Claude to reference ktp's HRT material and make it more accessible through translation, summarization, or rewording?
Absolutely not
The claudependent people throw stupid things like rewriting SentenceTransformers _but in rust_ as metrics
The average team doesn't need to rewrite SentenceTransformers _but in rust_ with Fable
GLM MoE sharded across a cluster is good enough to maintain terraform and astro sites, k8s clusters, etc.
GLM MoE sharded across a cluster is good enough to maintain terraform and astro sites, k8s clusters, etc.
August 20, 2026 at 7:49 PM
The average team doesn't need to rewrite SentenceTransformers _but in rust_ with Fable
GLM MoE sharded across a cluster is good enough to maintain terraform and astro sites, k8s clusters, etc.
GLM MoE sharded across a cluster is good enough to maintain terraform and astro sites, k8s clusters, etc.
I have some public bits and pieces for sqlite
This is bluesky -> embeddinggemma via SentenceTransformers in sqlite:
github.com/elijahsgh/sq...
This is rembed using ollama and embedding gemma:
github.com/elijahsgh/sq...
No classifier, though, but those two are enough for at least recommendations
This is bluesky -> embeddinggemma via SentenceTransformers in sqlite:
github.com/elijahsgh/sq...
This is rembed using ollama and embedding gemma:
github.com/elijahsgh/sq...
No classifier, though, but those two are enough for at least recommendations
January 15, 2026 at 5:36 AM
I have some public bits and pieces for sqlite
This is bluesky -> embeddinggemma via SentenceTransformers in sqlite:
github.com/elijahsgh/sq...
This is rembed using ollama and embedding gemma:
github.com/elijahsgh/sq...
No classifier, though, but those two are enough for at least recommendations
This is bluesky -> embeddinggemma via SentenceTransformers in sqlite:
github.com/elijahsgh/sq...
This is rembed using ollama and embedding gemma:
github.com/elijahsgh/sq...
No classifier, though, but those two are enough for at least recommendations
curious about methodology for semantic grouping here
> categorization uses SentenceTransformers
is the code open source / can you expand on this?
> categorization uses SentenceTransformers
is the code open source / can you expand on this?
February 20, 2026 at 9:30 PM
curious about methodology for semantic grouping here
> categorization uses SentenceTransformers
is the code open source / can you expand on this?
> categorization uses SentenceTransformers
is the code open source / can you expand on this?
We uniquely integrate deep-learning language models with traditional data to enhance the prediction of disease subtypes. 🧠 #DeepLearning #InnovativeResearch #languagemodels #sentenceTransformers #huggingface #ML #NLP
December 7, 2024 at 5:14 PM
We uniquely integrate deep-learning language models with traditional data to enhance the prediction of disease subtypes. 🧠 #DeepLearning #InnovativeResearch #languagemodels #sentenceTransformers #huggingface #ML #NLP
🔗 Learn more:
• Official website → sbert.net
• Original paper → aclanthology.org/D19-1410.pdf
• GitHub repository → github.com/UKPLab/sente...
(3/🧵)
• Official website → sbert.net
• Original paper → aclanthology.org/D19-1410.pdf
• GitHub repository → github.com/UKPLab/sente...
(3/🧵)
SentenceTransformers Documentation — Sentence Transformers documentation
sbert.net
October 22, 2025 at 2:08 PM
🔗 Learn more:
• Official website → sbert.net
• Original paper → aclanthology.org/D19-1410.pdf
• GitHub repository → github.com/UKPLab/sente...
(3/🧵)
• Official website → sbert.net
• Original paper → aclanthology.org/D19-1410.pdf
• GitHub repository → github.com/UKPLab/sente...
(3/🧵)
Hi everyone!🤗
I just released 𝗦𝗲𝗻𝗧𝗿𝗘𝘃, a python package that offers simple customizable #evaluation for text #retrieval performance of #SentenceTransformers embedders on PDFs.
More details on the LinkedIn post:
www.linkedin.com/posts/astra-...
And on the GitHub repo:
github.com/AstraBert/Se...
I just released 𝗦𝗲𝗻𝗧𝗿𝗘𝘃, a python package that offers simple customizable #evaluation for text #retrieval performance of #SentenceTransformers embedders on PDFs.
More details on the LinkedIn post:
www.linkedin.com/posts/astra-...
And on the GitHub repo:
github.com/AstraBert/Se...
November 25, 2024 at 11:17 AM
Hi everyone!🤗
I just released 𝗦𝗲𝗻𝗧𝗿𝗘𝘃, a python package that offers simple customizable #evaluation for text #retrieval performance of #SentenceTransformers embedders on PDFs.
More details on the LinkedIn post:
www.linkedin.com/posts/astra-...
And on the GitHub repo:
github.com/AstraBert/Se...
I just released 𝗦𝗲𝗻𝗧𝗿𝗘𝘃, a python package that offers simple customizable #evaluation for text #retrieval performance of #SentenceTransformers embedders on PDFs.
More details on the LinkedIn post:
www.linkedin.com/posts/astra-...
And on the GitHub repo:
github.com/AstraBert/Se...
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting Ни одного из этих слов в м...
#deepseek #llm #nlp #python #SentenceTransformers #seo #автоматизация #кластеризация #семантики #семантическое #ядро
Origin | Interest | Match
#deepseek #llm #nlp #python #SentenceTransformers #seo #автоматизация #кластеризация #семантики #семантическое #ядро
Origin | Interest | Match
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting
www.pvsm.ru
March 26, 2026 at 12:13 PM
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting Ни одного из этих слов в м...
#deepseek #llm #nlp #python #SentenceTransformers #seo #автоматизация #кластеризация #семантики #семантическое #ядро
Origin | Interest | Match
#deepseek #llm #nlp #python #SentenceTransformers #seo #автоматизация #кластеризация #семантики #семантическое #ядро
Origin | Interest | Match
😸 Loved this #PyconUS lightning talk:
"Find Better 🐱 Cat Emojis with your text!"
slides.takanory.net/slides/20260...
Takanori made emojis searchable by generating descriptions with Ollama+gemma4, then embedded with SentenceTransformers + multilingual-e5-base, and searched with ANN.
"Find Better 🐱 Cat Emojis with your text!"
slides.takanory.net/slides/20260...
Takanori made emojis searchable by generating descriptions with Ollama+gemma4, then embedded with SentenceTransformers + multilingual-e5-base, and searched with ANN.
Find Better 🐱 Cat Emojis with your text!
Takanori Suzuki PyCon US 2026 logo PyCon US 2026 / 2026 May 17[AD] PyCon JP 2026 🇯🇵: 2026.pycon.jp, 📅 2026 Aug 21(Fr)- 23(Su), ⛩️ Hiroshima, Japan, 📣 Call for Proposals, Call for Sponsors, Tickets ...
slides.takanory.net
May 17, 2026 at 3:56 PM
😸 Loved this #PyconUS lightning talk:
"Find Better 🐱 Cat Emojis with your text!"
slides.takanory.net/slides/20260...
Takanori made emojis searchable by generating descriptions with Ollama+gemma4, then embedded with SentenceTransformers + multilingual-e5-base, and searched with ANN.
"Find Better 🐱 Cat Emojis with your text!"
slides.takanory.net/slides/20260...
Takanori made emojis searchable by generating descriptions with Ollama+gemma4, then embedded with SentenceTransformers + multilingual-e5-base, and searched with ANN.
I was in particular looking for a way to make the theories practical and easy to both explain and execute.
For now, it seems that a combo of data gathering (apify, dataforseo), intelligence (pyton, sentencetransformers), and storage (pgvector) would be a solid & cheap infrastructure
For now, it seems that a combo of data gathering (apify, dataforseo), intelligence (pyton, sentencetransformers), and storage (pgvector) would be a solid & cheap infrastructure
February 5, 2026 at 7:16 AM
I was in particular looking for a way to make the theories practical and easy to both explain and execute.
For now, it seems that a combo of data gathering (apify, dataforseo), intelligence (pyton, sentencetransformers), and storage (pgvector) would be a solid & cheap infrastructure
For now, it seems that a combo of data gathering (apify, dataforseo), intelligence (pyton, sentencetransformers), and storage (pgvector) would be a solid & cheap infrastructure
A few days ago, I released the first alpha of FileChat — a local, read-only AI coding assistant for more conservative developers.
Some quick updates since then:
- Got a pull request from a third-party contributor (super motivating!)
- Replaced SentenceTransformers + PyTorch with ONNX Runtime → […]
Some quick updates since then:
- Got a pull request from a third-party contributor (super motivating!)
- Replaced SentenceTransformers + PyTorch with ONNX Runtime → […]
Original post on mastodon.social
mastodon.social
September 18, 2025 at 12:00 PM
A few days ago, I released the first alpha of FileChat — a local, read-only AI coding assistant for more conservative developers.
Some quick updates since then:
- Got a pull request from a third-party contributor (super motivating!)
- Replaced SentenceTransformers + PyTorch with ONNX Runtime → […]
Some quick updates since then:
- Got a pull request from a third-party contributor (super motivating!)
- Replaced SentenceTransformers + PyTorch with ONNX Runtime → […]
Not LDA, just k-means: and then describe each cluster based only on the core, not trying to capture all the variety. (Which makes the lack of distinction even more surprising, TBH).
September 23, 2023 at 1:09 AM
Not LDA, just k-means: and then describe each cluster based only on the core, not trying to capture all the variety. (Which makes the lack of distinction even more surprising, TBH).
Yooo, Sparse Neural Retrieval fi-na-lly getting the attention it deserves!
Kudos to @hf.co for making this 5th version of SentenceTransformers real, I really-really want to see what the HF community will craft with this 👀👀👀
🔥 huggingface.co/blog/train-s...
Kudos to @hf.co for making this 5th version of SentenceTransformers real, I really-really want to see what the HF community will craft with this 👀👀👀
🔥 huggingface.co/blog/train-s...
huggingface.co
July 1, 2025 at 2:20 PM
Yooo, Sparse Neural Retrieval fi-na-lly getting the attention it deserves!
Kudos to @hf.co for making this 5th version of SentenceTransformers real, I really-really want to see what the HF community will craft with this 👀👀👀
🔥 huggingface.co/blog/train-s...
Kudos to @hf.co for making this 5th version of SentenceTransformers real, I really-really want to see what the HF community will craft with this 👀👀👀
🔥 huggingface.co/blog/train-s...
the enthusiasm for summarization eludes me, particularly when LLMs have been one of those once-in-a-decade advances in search ranking (the first 10 years of TREC gave us... BM25)
You gotta think quantitatively, at least a little bit, to put that to work [1]
[1] sbert.net
You gotta think quantitatively, at least a little bit, to put that to work [1]
[1] sbert.net
SentenceTransformers Documentation — Sentence Transformers documentation
sbert.net
February 14, 2025 at 3:10 PM
the enthusiasm for summarization eludes me, particularly when LLMs have been one of those once-in-a-decade advances in search ranking (the first 10 years of TREC gave us... BM25)
You gotta think quantitatively, at least a little bit, to put that to work [1]
[1] sbert.net
You gotta think quantitatively, at least a little bit, to put that to work [1]
[1] sbert.net
SentenceTransformersがHF🤗にJoinしたそう。（なぜか）すでにSBERTはHF傘下と思い込んでいたので，昨日Joinしたというアナウンスにむしろ驚いてる。
huggingface.co/blog/sentenc...
huggingface.co/blog/sentenc...
Sentence Transformers is joining Hugging Face!
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
October 23, 2025 at 6:19 AM
SentenceTransformersがHF🤗にJoinしたそう。（なぜか）すでにSBERTはHF傘下と思い込んでいたので，昨日Joinしたというアナウンスにむしろ驚いてる。
huggingface.co/blog/sentenc...
huggingface.co/blog/sentenc...
📰 Read the full announcements:
TU Darmstadt Press Release → www.tu-darmstadt.de/universitaet...
Hugging Face Blog Post → huggingface.co/blog/sentenc...
(4/4)
#UKPLab #HuggingFace #SentenceTransformers #NLP #AIresearch #OpenSource 🚀
TU Darmstadt Press Release → www.tu-darmstadt.de/universitaet...
Hugging Face Blog Post → huggingface.co/blog/sentenc...
(4/4)
#UKPLab #HuggingFace #SentenceTransformers #NLP #AIresearch #OpenSource 🚀
So versteht KI ganze Sätze
Das Ubiquitous Knowledge Processing (UKP) Lab der Technischen Universität Darmstadt hat die Wartung und Weiterentwicklung von Sentence Transformers, einer der weltweit meistgenutzten Open-Source-Bibli...
www.tu-darmstadt.de
October 22, 2025 at 2:08 PM
📰 Read the full announcements:
TU Darmstadt Press Release → www.tu-darmstadt.de/universitaet...
Hugging Face Blog Post → huggingface.co/blog/sentenc...
(4/4)
#UKPLab #HuggingFace #SentenceTransformers #NLP #AIresearch #OpenSource 🚀
TU Darmstadt Press Release → www.tu-darmstadt.de/universitaet...
Hugging Face Blog Post → huggingface.co/blog/sentenc...
(4/4)
#UKPLab #HuggingFace #SentenceTransformers #NLP #AIresearch #OpenSource 🚀
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting Началось с того что мне на...
#SEO #кластеризация #семантики #Python #NLP #LLM #DeepSeek #Яндекс #Директ #SentenceTransformers #семантическое
Origin | Interest | Match
#SEO #кластеризация #семантики #Python #NLP #LLM #DeepSeek #Яндекс #Директ #SentenceTransformers #семантическое
Origin | Interest | Match
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting
Ни одного из этих слов в моих планах не было. Я просто задолбался вручную таскать ключи из Wordstat в Excel. Версия 1: лишь бы не копировать руками Знакомая ситуация: открываешь Wordstat, вводишь...
habr.com
March 26, 2026 at 11:21 AM
Хотел перестать копировать из Wordstat. Получилась мультиагентная система с Ensemble Voting Началось с того что мне на...
#SEO #кластеризация #семантики #Python #NLP #LLM #DeepSeek #Яндекс #Директ #SentenceTransformers #семантическое
Origin | Interest | Match
#SEO #кластеризация #семантики #Python #NLP #LLM #DeepSeek #Яндекс #Директ #SentenceTransformers #семантическое
Origin | Interest | Match
Wrote down some learnings from finetuning embedding models: fixing a SentenceTransformers batch sampler bug, building cached supervised contrastive loss, and running hyperparameter searches over semantic knobs that compile into GPU-specific execution plans.
www.mrloh.io/blog/2026-05...
www.mrloh.io/blog/2026-05...
Automating GPU-Aware Finetuning for Embedding Models
How chasing a broken batch sampler turned into rewriting our embedding finetuning stack around cached supervised contrastive loss and a compiler that turns training intent into hardware-aware executio...
www.mrloh.io
May 21, 2026 at 3:45 PM
Wrote down some learnings from finetuning embedding models: fixing a SentenceTransformers batch sampler bug, building cached supervised contrastive loss, and running hyperparameter searches over semantic knobs that compile into GPU-specific execution plans.
www.mrloh.io/blog/2026-05...
www.mrloh.io/blog/2026-05...