viidfedenwad

viidfedenwad

AI & ML interests

None yet

Recent Activity

reacted to seawolf2357's post with ā¤ļø 1 day ago
šŸš€ Just Found an Interesting New Leaderboard for Medical AI Evaluation! I recently stumbled upon a medical domain-specific FACTS Grounding leaderboard on Hugging Face, and the approach to evaluating AI accuracy in medical contexts is quite impressive, so I thought I'd share. šŸ“Š What is FACTS Grounding? It's originally a benchmark developed by Google DeepMind that measures how well LLMs generate answers based solely on provided documents. What's cool about this medical-focused version is that it's designed to test even small open-source models. šŸ„ Medical Domain Version Features 236 medical examples: Extracted from the original 860 examples Tests small models like Qwen 3 1.7B: Great for resource-constrained environments Uses Gemini 1.5 Flash for evaluation: Simplified to a single judge model šŸ“ˆ The Evaluation Method is Pretty Neat Grounding Score: Are all claims in the response supported by the provided document? Quality Score: Does it properly answer the user's question? Combined Score: Did it pass both checks? Since medical information requires extreme accuracy, this thorough verification approach makes a lot of sense. šŸ”— Check It Out Yourself The actual leaderboard: https://huggingface.co/spaces/MaziyarPanahi/FACTS-Leaderboard šŸ’­ My thoughts: As medical AI continues to evolve, evaluation tools like this are becoming increasingly important. The fact that it can test smaller models is particularly helpful for the open-source community!
reacted to seawolf2357's post with šŸ¤— 1 day ago
šŸš€ Just Found an Interesting New Leaderboard for Medical AI Evaluation! I recently stumbled upon a medical domain-specific FACTS Grounding leaderboard on Hugging Face, and the approach to evaluating AI accuracy in medical contexts is quite impressive, so I thought I'd share. šŸ“Š What is FACTS Grounding? It's originally a benchmark developed by Google DeepMind that measures how well LLMs generate answers based solely on provided documents. What's cool about this medical-focused version is that it's designed to test even small open-source models. šŸ„ Medical Domain Version Features 236 medical examples: Extracted from the original 860 examples Tests small models like Qwen 3 1.7B: Great for resource-constrained environments Uses Gemini 1.5 Flash for evaluation: Simplified to a single judge model šŸ“ˆ The Evaluation Method is Pretty Neat Grounding Score: Are all claims in the response supported by the provided document? Quality Score: Does it properly answer the user's question? Combined Score: Did it pass both checks? Since medical information requires extreme accuracy, this thorough verification approach makes a lot of sense. šŸ”— Check It Out Yourself The actual leaderboard: https://huggingface.co/spaces/MaziyarPanahi/FACTS-Leaderboard šŸ’­ My thoughts: As medical AI continues to evolve, evaluation tools like this are becoming increasingly important. The fact that it can test smaller models is particularly helpful for the open-source community!
reacted to seawolf2357's post with šŸ”„ 1 day ago
šŸš€ Just Found an Interesting New Leaderboard for Medical AI Evaluation! I recently stumbled upon a medical domain-specific FACTS Grounding leaderboard on Hugging Face, and the approach to evaluating AI accuracy in medical contexts is quite impressive, so I thought I'd share. šŸ“Š What is FACTS Grounding? It's originally a benchmark developed by Google DeepMind that measures how well LLMs generate answers based solely on provided documents. What's cool about this medical-focused version is that it's designed to test even small open-source models. šŸ„ Medical Domain Version Features 236 medical examples: Extracted from the original 860 examples Tests small models like Qwen 3 1.7B: Great for resource-constrained environments Uses Gemini 1.5 Flash for evaluation: Simplified to a single judge model šŸ“ˆ The Evaluation Method is Pretty Neat Grounding Score: Are all claims in the response supported by the provided document? Quality Score: Does it properly answer the user's question? Combined Score: Did it pass both checks? Since medical information requires extreme accuracy, this thorough verification approach makes a lot of sense. šŸ”— Check It Out Yourself The actual leaderboard: https://huggingface.co/spaces/MaziyarPanahi/FACTS-Leaderboard šŸ’­ My thoughts: As medical AI continues to evolve, evaluation tools like this are becoming increasingly important. The fact that it can test smaller models is particularly helpful for the open-source community!
View all activity

Organizations

None yet

viidfedenwad's activity

reacted to seawolf2357's post with ā¤ļøšŸ¤—šŸ”„šŸ‘ 1 day ago
view post
Post
959
šŸš€ Just Found an Interesting New Leaderboard for Medical AI Evaluation!

I recently stumbled upon a medical domain-specific FACTS Grounding leaderboard on Hugging Face, and the approach to evaluating AI accuracy in medical contexts is quite impressive, so I thought I'd share.

šŸ“Š What is FACTS Grounding?
It's originally a benchmark developed by Google DeepMind that measures how well LLMs generate answers based solely on provided documents. What's cool about this medical-focused version is that it's designed to test even small open-source models.

šŸ„ Medical Domain Version Features

236 medical examples: Extracted from the original 860 examples
Tests small models like Qwen 3 1.7B: Great for resource-constrained environments
Uses Gemini 1.5 Flash for evaluation: Simplified to a single judge model

šŸ“ˆ The Evaluation Method is Pretty Neat

Grounding Score: Are all claims in the response supported by the provided document?
Quality Score: Does it properly answer the user's question?
Combined Score: Did it pass both checks?

Since medical information requires extreme accuracy, this thorough verification approach makes a lot of sense.
šŸ”— Check It Out Yourself

The actual leaderboard: MaziyarPanahi/FACTS-Leaderboard

šŸ’­ My thoughts: As medical AI continues to evolve, evaluation tools like this are becoming increasingly important. The fact that it can test smaller models is particularly helpful for the open-source community!
reacted to ginipick's post with šŸš€ 5 days ago
view post
Post
4022
šŸŽØ FLUX VIDEO Generation - All-in-One AI Image/Video/Audio Generator

šŸš€ Introduction
FLUX VIDEO Generation is an all-in-one AI creative tool that generates images, videos, and audio from text prompts, powered by NVIDIA H100 GPU for lightning-fast processing!

ginigen/Flux-VIDEO

✨ Key Features
1ļøāƒ£ Text → Image → Video šŸ–¼ļøāž”ļøšŸŽ¬

Generate high-quality images from Korean/English prompts
Transform still images into natural motion videos
Multiple size presets (Instagram, YouTube, Facebook, etc.)
Demo: 1-4 seconds / Full version: up to 60 seconds

2ļøāƒ£ Image Aspect Ratio Change šŸŽ­

Freely adjust image aspect ratios
Expand images with outpainting technology
5 alignment options (Center, Left, Right, Top, Bottom)
Real-time preview functionality

3ļøāƒ£ Video + Audio Generation šŸŽµ

Add AI-generated audio to videos
Korean prompt support (auto-translation)
Context-aware sound generation
Powered by MMAudio technology

šŸ› ļø Tech Stack

Image Generation: FLUX, Stable Diffusion XL
Video Generation: TeaCache optimization
Audio Generation: MMAudio (44kHz high-quality)
Outpainting: ControlNet Union
Infrastructure: NVIDIA H100 GPU for ultra-fast generation

šŸ’” How to Use

Select your desired tab
Enter your prompt (Korean/English supported!)
Adjust settings
Click generate button

šŸŽÆ Use Cases

šŸ“± Social media content creation
šŸŽ„ YouTube Shorts/Reels
šŸ“Š Presentation materials
šŸŽØ Creative artwork
šŸŽµ Background sound generation
  • 1 reply
Ā·
reacted to ginipick's post with šŸš€ 5 days ago
view post
Post
3576
šŸŽØ AI Hairstyle Changer - Transform with 93 Styles! šŸ’‡ā€ā™€ļøāœØ

šŸš€ Introduction
Experience 93 different hairstyles and 29 hair colors in real-time with your uploaded photo!
Transform your look instantly with this AI-powered Gradio web app.


✨ Key Features

šŸ“ø Simple 3 Steps
Upload Photo - Upload a front-facing photo
Select Style - Choose from 93 hairstyles
Pick Color - Click your desired color from 29 color palette options


šŸ’« Diverse Hairstyles (93 types)

šŸŽÆ Short Cuts: Pixie Cut, Bob, Lob, Crew Cut, Undercut
🌊 Waves: Soft Waves, Hollywood Waves, Finger Waves
šŸŽ€ Braids: French Braid, Box Braids, Fishtail Braid, Cornrows
šŸ‘‘ Updos: Chignon, Messy Bun, Top Knot, French Twist
🌈 Special Styles: Space Buns, Dreadlocks, Mohawk, Beehive

šŸŽØ Hair Color Palette (29 colors)

šŸ¤Ž Natural Colors: Black, Browns, Blonde variations
ā¤ļø Red Tones: Red, Auburn, Copper, Burgundy
šŸ’œ Fashion Colors: Blue, Purple, Pink, Green, Rose Gold
⚪ Cool Tones: Silver, Ash Blonde, Titanium

🌟 Key Advantages

⚔ Fast Processing: Get results in just 10-30 seconds
šŸŽÆ High Accuracy: Natural-looking transformations with AI technology
šŸ’Ž Professional Quality: High-resolution output suitable for social media
šŸ”„ Unlimited Trials: Try as many combinations as you want
šŸ“± User-Friendly: Intuitive interface with visual color palette


šŸ’” Perfect For

šŸ’ˆ Salon Consultations: Show clients potential new looks before cutting
šŸ›ļø Personal Styling: Experiment before making a big change
šŸŽ­ Entertainment: Fun transformations for social media content
šŸŽ¬ Creative Projects: Character design and visualization
šŸ‘— Fashion Industry: Match hairstyles with outfits and makeup
šŸ“ø Photography: Pre-visualization for photoshoots

LINK: ginipick/Change-Hair
Ā·
reacted to openfree's post with šŸš€šŸ‘€ 7 days ago
view post
Post
2691
šŸŽ™ļø Voice Clone AI Podcast Generator: Create Emotionally Rich Podcasts with Your Own Voice!

šŸš€ Project Introduction
Hello! Today we're excited to introduce an AI-powered solo podcast generator that creates high-quality voice cloning with authentic emotional expression.
Transform any PDF document, web URL, or keyword into a professional podcast with just a few clicks! šŸ“šāž”ļøšŸŽ§

VIDraft/Voice-Clone-Podcast

✨ Key Features
1. šŸŽÆ Multiple Input Methods

URL: Simply paste any blog or article link
PDF: Upload research papers or documents directly
Keyword: Enter a topic and AI searches for the latest information to create content

2. šŸŽ­ Emotionally Expressive Voice Cloning
Powered by Chatterbox TTS:

šŸŽ¤ Voice Cloning: Learn and replicate your unique voice perfectly
šŸ“¢ Natural intonation and emotional expression
🌊 Customizable emotion intensity with Exaggeration control
⚔ Seamless handling of long texts with automatic chunking

3. šŸ¤– State-of-the-Art LLM Script Generation

Professional-grade English dialogue using Private-BitSix-Mistral
12 natural conversational exchanges
Real-time web search integration for up-to-date information
Fully editable generated scripts! āœļø

šŸ’” Use Cases
šŸ“– Educational Content

Transform complex research papers into easy-to-understand podcasts
Create English learning materials in your own voice

šŸ“° News & Information

Convert international articles into engaging audio content
Produce global trend analysis podcasts

šŸŽØ Creative Content

Tell stories in English with your own voice
Build your global personal brand with custom audio content

šŸ› ļø Tech Stack
🧠 LLM: Llama CPP + Private-BitSix-Mistral
šŸ—£ļø TTS: Chatterbox (Voice Cloning & Emotional Expression)
šŸ” Search: Brave Search API
šŸ“„ Document Processing: LangChain + PyPDF
šŸ–„ļø Interface: Gradio
šŸŽ‰ What Makes Us Special

šŸŽ¤ Voice Cloning: Perfect voice replication from just a short audio sample
😊 Emotion Contro šŸ“ Unlimited Length šŸ”„ Real-time Updates
  • 1 reply
Ā·
reacted to openfree's post with šŸ”„ 8 days ago
view post
Post
2691
šŸŽ™ļø Voice Clone AI Podcast Generator: Create Emotionally Rich Podcasts with Your Own Voice!

šŸš€ Project Introduction
Hello! Today we're excited to introduce an AI-powered solo podcast generator that creates high-quality voice cloning with authentic emotional expression.
Transform any PDF document, web URL, or keyword into a professional podcast with just a few clicks! šŸ“šāž”ļøšŸŽ§

VIDraft/Voice-Clone-Podcast

✨ Key Features
1. šŸŽÆ Multiple Input Methods

URL: Simply paste any blog or article link
PDF: Upload research papers or documents directly
Keyword: Enter a topic and AI searches for the latest information to create content

2. šŸŽ­ Emotionally Expressive Voice Cloning
Powered by Chatterbox TTS:

šŸŽ¤ Voice Cloning: Learn and replicate your unique voice perfectly
šŸ“¢ Natural intonation and emotional expression
🌊 Customizable emotion intensity with Exaggeration control
⚔ Seamless handling of long texts with automatic chunking

3. šŸ¤– State-of-the-Art LLM Script Generation

Professional-grade English dialogue using Private-BitSix-Mistral
12 natural conversational exchanges
Real-time web search integration for up-to-date information
Fully editable generated scripts! āœļø

šŸ’” Use Cases
šŸ“– Educational Content

Transform complex research papers into easy-to-understand podcasts
Create English learning materials in your own voice

šŸ“° News & Information

Convert international articles into engaging audio content
Produce global trend analysis podcasts

šŸŽØ Creative Content

Tell stories in English with your own voice
Build your global personal brand with custom audio content

šŸ› ļø Tech Stack
🧠 LLM: Llama CPP + Private-BitSix-Mistral
šŸ—£ļø TTS: Chatterbox (Voice Cloning & Emotional Expression)
šŸ” Search: Brave Search API
šŸ“„ Document Processing: LangChain + PyPDF
šŸ–„ļø Interface: Gradio
šŸŽ‰ What Makes Us Special

šŸŽ¤ Voice Cloning: Perfect voice replication from just a short audio sample
😊 Emotion Contro šŸ“ Unlimited Length šŸ”„ Real-time Updates
  • 1 reply
Ā·
replied to clem's post 8 days ago
view reply

Wow, I want to buy it right now. Where can I buy it?

reacted to clem's post with šŸ”„ 8 days ago
view post
Post
5390
Today, we're unveiling two new open-source AI robots! HopeJR for $3,000 & Reachy Mini for $300 šŸ¤–šŸ¤–šŸ¤–

Let's go open-source AI robotics!
Ā·
reacted to openfree's post with šŸ”„ 10 days ago
view post
Post
2567
🧠 AI Brand Naming with 15 Specialized Theories

šŸŽÆ Core Features
15 Expert Theories for professional brand naming
Bilingual Support Korean/English for global brands
Unified Evaluation System creativity/memorability/relevance scores
Real-time Visualization theory-specific custom designs

openfree/Naming

šŸ”¬ Applied Theories
Cognitive Theories (4)
🟦 Square Theory - Semantic square structure with 4-word relationships
šŸ”Š Sound Symbolism - Psychological connections between phonemes and meaning
🧠 Cognitive Load - Minimized processing for instant recognition
šŸ‘ļø Gestalt Theory - Perceptual principles where whole exceeds parts

Creative Theories (3)
šŸ”€ Conceptual Blending - Merging concepts to create new meanings
šŸ”§ SCAMPER Method - 7 creative transformation techniques
🌿 Biomimicry - Nature-inspired wisdom from 3.8 billion years of evolution

Strategic Theories (2)
āœ… Jobs-to-be-Done - Customer-centric problem-solving focus
šŸ’­ Design Thinking - Human-centered innovation methodology

Cultural Theories (3)
šŸŽ­ Jung's Archetype - 12 universal archetypes for emotional connection
🌐 Linguistic Relativity - Cross-cultural thinking patterns consideration
🧬 Memetics - Cultural transmission and evolutionary potential

Differentiation Theories (3)
⚔ Von Restorff Effect - Uniqueness for 30x better recall
šŸŽØ Color Psychology - Emotional associations and color meanings
šŸŒ Network Effects - Value maximization through network structures

šŸ’« Special Features
Each theory provides unique visualizations and customized analysis:

Square Theory → 4-corner relationship diagram
Blending → Concept fusion flowchart
Color → Interactive color palette display
Theory-specific insights for each approach

šŸŽØ Output Information
Core: Brand name, slogan, values, emotions, personality
Visual: Colors, concepts, typography styles
Linguistic: Pronunciation, etymology, global adaptability
Strategic: Differentiation, positioning, growth potential
Theory-specific...