Tag
hugging-face
21 posts · all tags
23 Cheshvan 5787
Can image models spell the labels? A first look at my Tech Graphics Benchmark
Limited evaluation: 95 images from five cheap image models, 12 annotated so far. In this run, 10 of 11 had every label spelled exactly; the problems were composition, stretching and style. Not a verdict on any model.
23 Cheshvan 5787
Datasets for fine-tuning speech recognition on your own voice
A guide to my public Hugging Face datasets of read-aloud sentences for fine-tuning speech recognition: tech, Linux, smart home, AI agents, homophones, Hebrew code-switching and voice notes. What each holds and which to use.
23 Cheshvan 5787
My AI prompt libraries, and which one to use
A guide to my open-source system prompt library and its siblings: prompts for NotebookLM, Open WebUI, Roo Code and ChatGPT, tools for writing prompts, and which of the twelve repositories and Hugging Face datasets to use for what.
23 Cheshvan 5787
Contextual Sentence Creator: a 3B model that writes ASR training sentences
A Granite 4.2 3B fine-tune that writes natural sentences using a given term, for recording speech recognition training data. 95.7% of sentences pass the acceptance check; GGUF build included.
23 Cheshvan 5787
Fritter: a Parakeet fine-tune for English tech dictation
Fritter is my Parakeet TDT 0.6B v2 fine-tune for English dictation without Hebrew: tech terms, numbers and brand names. Tech WER falls from 7.53% to 4.01% on my voice. Browser demo and GGUF builds.
23 Cheshvan 5787
Hey Yossi: a custom wake word for Home Assistant
Hey Yossi is a custom wake word for Home Assistant, trained on real recordings, in two builds: openWakeWord for the server and Pi, microWakeWord for ESPHome, Voice PE and Android. 95% recall in testing.
23 Cheshvan 5787
S1-ASR-Cleanup: cleaning up dictation with a fine-tuned S1-mini
A fine-tune of S1-mini by Superwhisper that cleans speech-to-text dictation: paragraphs, dates and money as digits, contractions and fillers handled. 0.6B, with GGUF builds for llama.cpp and phones.
23 Cheshvan 5787 Update
Update: a 32 MB model for ASR paragraph breaks
MiniLM-L12-ASR-Paragraph-Breaks replaces my Gemma 3 270M paragraph model: break F1 0.85 against 0.78, 32 MB instead of 437 MB, and it cannot change a word of the transcript.
23 Cheshvan 5787
Agent Task Break Detector: knowing when to start a fresh session
A small ModernBERT model that checks every message in an AI agent session and says whether it is a new task a fresh session could handle. Runs locally on CPU; v0.1, trained on synthetic data.
16 Cheshvan 5787
Using AI to decide my vote in Israel's 2026 election
How I turned a 42-minute voice note into an AI vote decision for Israel's 2026 election: the method in diagrams, my result, and a step-by-step guide to doing it yourself with any AI chat app. No list has a real tenant-protection policy.
11 Cheshvan 5787
Forecasting Israel's election for two cents: decision models vs the markets
Two decision models answered 39 questions on Israel's 27 October election from a frozen pack of polls, lists and news, with and without market prices. It cost $0.016, and adding market prices flipped the forecast.
11 Cheshvan 5787
I had a conversation with my own voice clone
I fine-tuned a text-to-speech model on 88 minutes of my voice, then had a ten-minute conversation with it. My lines are real phone recordings; its lines are generated. Listen, and see how it was made.
10 Cheshvan 5787
Adding paragraph breaks to speech-to-text output with Gemma 3 270M
A Gemma 3 270M fine-tune that adds paragraph breaks to flat speech-to-text output and changes nothing else. Small enough for on-device use, with an ONNX export. Break F1 0.78 on 26 held-out memos.
10 Cheshvan 5787
Bolding the key lines of a dictated email with Gemma 3 270M
A preliminary Gemma 3 270M fine-tune that bolds the key lines of a dictated email (the asks, the purpose, the key facts) and changes nothing else. Small enough for on-device use. Word F1 0.62 on six held-out emails.
10 Cheshvan 5787
Proximate Pseudonymizer: a small model that swaps identifying details for close substitutes
A proof-of-concept Qwen3-0.6B fine-tune that rewrites text with names, places and dates swapped for close, consistent substitutes (Daniel to David, Jerusalem to Tel Aviv). Runs on CPU. Still leaks 9.3% of details, so review every output.
9 Cheshvan 5787
Boxty v1.0: English dictation without the Hebrew
Boxty is the no-Hebrew sibling of my Jachnun code-switching model: Parakeet fine-tuned on my voice for plain English and tech dictation. Tech WER drops from 2.51% (stock) to 1.35%, and general English holds.
9 Cheshvan 5787
Do LLMs push back on conspiracy believers? A small first look
A small, hand-graded first look at my conspiracy-endorsement eval: 7 prompts, 7 models, one sample each. In this run two uncensored models helped a believer and an invented theory tripped up two more. Not a verdict on any model.
4 Cheshvan 5787 Update
inventory-spec-extractor: a 34 MB model for reading product spec lines
A small token-classification model that pulls voltages, connectors, materials and 22 other attributes out of one line of a product spec. Runs on CPU in about 20 ms per line. F1 0.85 against a regex baseline's 0.58.
4 Cheshvan 5787
Can speech-to-text spell ‘makolet’? I tested seven models on Hebrish
English speakers in Israel talk Hebrish, and dictation software mangles the Hebrew words. I built an open benchmark to measure it: my fine-tune gets two-thirds of them right, off-the-shelf models a third to under half.
4 Cheshvan 5787
My Hugging Face models and datasets, by theme (October 2026)
A point-in-time snapshot of the 16 models and 72 datasets on my Hugging Face profile, grouped by theme, from Hebrew-English code-switching to Jerusalem photography.
4 Cheshvan 5787
Teaching a podcast to say makolet: a classifier for Hebrew hiding in English
I released a small model on Hugging Face that finds romanised Hebrew words inside English text, so TTS voices like Chatterbox can stop reading them as English. Built for My Weird Prompts.