POCs - a danielrosehill Collection

danielrosehill 's Collections

Paragraph Inference

Unified embedding models

Agent Accessory Models

Email Key Line Bolding

Paragraph Inference Postprocessing (V1) - Gemma 3

Audio Training Datasets (Sentences)

Hebrew audio datasets

Image Task Training

Israel Context Datasets

Open Data Packages

Evals-Runs-Datasets

Deepseek 4.1 Flash point releases

Parakeet V2 - English-Hebrew Codeswitching Finetune

English Hebrew Codeswitching - Datasets And Training

English-Hebrew Codeswitching - Models

Inventory Accessories

My Models - Classifiers

Code Switching (English And Hebrew) - All Resources

Fine Tune Candidates 1026

Baby Cry Detection

Chatterbox Multilingual

Token classifiers

ASR Models 0926

Pidgeon Detection NVR

Small ASR Models

Small Text Generation

IVM And Text Restoration

Punctuation Restoration- English

Interesting Spaces

On-Device ASR - One Plus CPH2493 (Nord 3)

DeepSeek V4.1 Derivatives

On Device Models

Use Case Diversity

Decision Making Support

Deepseek V4 Derivs

STT Cleanup (Text Postprocessing)

Android STT - Favorites

Israel Open Data

Acronym Identification

LLMS-Im-Testing

Hebrew Puncutation Restoration

Hebrew Sentiment Classification Models

English Hebrew Translation

Hebrew OCR Models

Hebrew Diacritic Restoration Models

Hebrew ASR (And Audio To Text)

Streaming-Speech-To-Text

Agentic Code Gen 301225

MWP-TTS-Candidates

Video Understanding

Image Evaluations

Evaluation Datasets

Audio Understanding Datasets

Voxtral Originals (Mistral)

Audio Multimodal Models

My Public Audio Datasets

ASR Beyond Whisper

Model Structures

Whisper Hebrish

My-ASR-Finetunes

ASR Benchmarking

Whisper Fine Tunes

API Price Comparisons

Video background removal

Architecture Related Models

STT Fine Tune Resources

Concept Outlines

STT Evaluations

Whisper Base + variants

Whisper variants

Voice Modality Apps

Entertainment Recommendations

Worlds (3D, Games)

Proofs of Concept

Background Noise Removal

Project Indexes

Real Time Video To Video

Voice Enhancement

Context Utilities

Voice Prompting

Resume Utilities

Multi LLM Experiments

Geolocation Utilities

Image Generation

Object Detection

Text Processing Utilities

Developer Utilities

Data Visualization

Hugging Face Utilities

Subtitle generation

Video editing utilities

Audio editing utilities

Prompt engineering

Interesting ideas

Data Processing Utilities

Background Removal

Image captioning

Video Generation Quants

TTS With Dialog Support

OCR & Document Processing

Fast video generation

Long speech synthesis

Agentic code generation capable

Browser use capable

Code Generation Models

Video Generation

General LLM Quants

Reasoning Models

Embedding Models

Instructional LLMs

Image Generation Models

Image Generation Quants

Local model collection

Vision Language Models

My Image Datasets

Text Transformation

Reference / Lookup Datasets

Sustainability Projects

Voice Note Audio And Training

My System Prompt Collections

Israel Photo Galleries

3D Human Digital Humans

Generative-AI-Favorites

Single Shot Image To Image (Reference)

Speech To Text (STT)

Text To Speech (TTS)

Image To Video (No Audio)

Music Generation

Image Editing Utilities

Character-Generation

Global Value Factor Database (GVFD) - Visualisation And Data

Text Reformatting Apps

Hebrew AI Spaces

Hebrew datasets

Fav-Code-Generation-Models

Hebrew LLMs + Text Prediction

LLM-Experiments

POCs

updated 7 days ago

Early stage proof of concept models that require more training data, training, or rearchitecting.

danielrosehill/Gemma-3-270M-Email-Key-Line-Bolding

Text Generation • 0.3B • Updated 7 days ago • 217

Note ASR pipeline component for bolding (or highlighting) key text (asks, details) in lengthy emails and ASR generations
danielrosehill/Proximate-Pseudonymizer-0.6B

Text Generation • 0.6B • Updated 7 days ago • 219

Note Text in, text out obfuscation intended for use-cases in whistleblowing and identity protection (rather than classic PII detection)