Lightricks/LTX-2.3-22b-IC-LoRA-Motion-Track-Control
Any-to-Any • Updated • 23
Geo Maciolek PRO
GeoMaciolek
·
AI & ML interests
None yet
Recent Activity
new activity about 8 hours ago
mradermacher/model_requests:darkc0de/Mistral-Small-4-119B-2603-heretic liked a model about 8 hours ago
Lightricks/LTX-2.3-22b-IC-LoRA-Motion-Track-Control liked a model about 10 hours ago
fal/Qwen-Image-Edit-2511-Multiple-Angles-LoRAOrganizations
None yet
darkc0de/Mistral-Small-4-119B-2603-heretic
1
#2049 opened about 8 hours ago
by GeoMaciolek
liked a
model about 8 hours ago
liked a
model about 10 hours ago
Suggestion
👀 1
2
#600 opened 3 days ago
by 
drmcbride
Slow inference on rtx 3090
5
#16 opened 1 day ago
by 
Blakus
liked a
model 7 days ago
reactedto SeaWolf-AI's post with 👍 9 days ago
Post
8143
🚀 Introducing MARL — Runtime Middleware That Reduces LLM Hallucination Without Fine-Tuning
Now available on PyPI · GitHub · ClawHub · HuggingFace
AI models sense they could be wrong, but they can't actually fix what's broken.
🤗 Live A/B test: VIDraft/MARL
We evaluated 9 SOTA models (GPT-5.2, Claude Opus 4.6, Gemini 3 Pro, etc.) across 1,800 assessments in FINAL Bench and found a 39.2%p gap between "recognizing potential errors (MA=0.694)" and "actually finding and fixing them (ER=0.302)."
MARL (Model-Agnostic Runtime Middleware for LLMs) was built to close this metacognitive gap. It decomposes a single LLM call into a 5-stage expert pipeline (Hypothesis → Solver → Auditor → Adversarial Verifier → Synthesizer), transforming "answer in one shot" into "think, doubt, correct, and rewrite."
No weight modification — works instantly with GPT-5.4, Claude, Gemini, Llama, or any OpenAI API-compatible LLM by changing one line: base_url. Ships with 9 domain-specific emergence engines (invention, pharma, genomics, chemistry, ecology, law, and more — 5,538 expert data items) activated by a simple tag like model="gpt-5.4::pharma".
pip install marl-middleware
MARL is also officially registered on ClawHub, the skill marketplace of OpenClaw — an AI agent platform with 260K+ developers and 3,200+ skills. It's the first middleware in the Reasoning Enhancement category. One command — clawhub install marl-middleware — gives your AI agent a metacognition upgrade.
📝 Technical deep dive: https://huggingface.co/blog/FINAL-Bench/marl-middleware
📦 PyPI: https://pypi.org/project/marl-middleware/
🐙 GitHub: https://github.com/Vidraft/MARL
🦀 ClawHub: https://clawhub.ai/Cutechicken99/marl-middleware
#MARL #LLM #Hallucination #Metacognition #MultiAgent #AIMiddleware #FINALBench #OpenClaw #ClawHub #PyPI #AGI #HuggingFace #ReasoningAI #SelfCorrection #GlassBoxAI
Now available on PyPI · GitHub · ClawHub · HuggingFace
AI models sense they could be wrong, but they can't actually fix what's broken.
🤗 Live A/B test: VIDraft/MARL
We evaluated 9 SOTA models (GPT-5.2, Claude Opus 4.6, Gemini 3 Pro, etc.) across 1,800 assessments in FINAL Bench and found a 39.2%p gap between "recognizing potential errors (MA=0.694)" and "actually finding and fixing them (ER=0.302)."
MARL (Model-Agnostic Runtime Middleware for LLMs) was built to close this metacognitive gap. It decomposes a single LLM call into a 5-stage expert pipeline (Hypothesis → Solver → Auditor → Adversarial Verifier → Synthesizer), transforming "answer in one shot" into "think, doubt, correct, and rewrite."
No weight modification — works instantly with GPT-5.4, Claude, Gemini, Llama, or any OpenAI API-compatible LLM by changing one line: base_url. Ships with 9 domain-specific emergence engines (invention, pharma, genomics, chemistry, ecology, law, and more — 5,538 expert data items) activated by a simple tag like model="gpt-5.4::pharma".
pip install marl-middleware
MARL is also officially registered on ClawHub, the skill marketplace of OpenClaw — an AI agent platform with 260K+ developers and 3,200+ skills. It's the first middleware in the Reasoning Enhancement category. One command — clawhub install marl-middleware — gives your AI agent a metacognition upgrade.
📝 Technical deep dive: https://huggingface.co/blog/FINAL-Bench/marl-middleware
📦 PyPI: https://pypi.org/project/marl-middleware/
🐙 GitHub: https://github.com/Vidraft/MARL
🦀 ClawHub: https://clawhub.ai/Cutechicken99/marl-middleware
#MARL #LLM #Hallucination #Metacognition #MultiAgent #AIMiddleware #FINALBench #OpenClaw #ClawHub #PyPI #AGI #HuggingFace #ReasoningAI #SelfCorrection #GlassBoxAI
Best ever
4
#1 opened 13 days ago
by 
Sebas28
UGI Leaderboard Writin benchmark
4
#2 opened 5 months ago
by 
s1arsky
Feature (kinda) - A couple of Quant Comparisons
👀 1
2
#581 opened 15 days ago
by GeoMaciolek
ST settings ?
1
#1 opened about 1 month ago
by 
edreces
liked 3
models 15 days ago
sophosympatheia/Magistry-24B-v1.0 https://huggingface.co/sophosympatheia/Magistry-24B-v1.0
1
#1932 opened 17 days ago
by GeoMaciolek
NoesisLab/Kai-30B-Instruct https://huggingface.co/NoesisLab/Kai-30B-Instruct
1
#1931 opened 17 days ago
by GeoMaciolek
Duplicate entry for 'YanLabs/gemma3-27b-it-abliterated-normpreserve'
4
#570 opened 19 days ago
by GeoMaciolek
liked 2
models 18 days ago
Nice placement on the UGI Leaderboards - #5 for writing, for 24B and under.
1
#1 opened 18 days ago
by GeoMaciolek