What If the Adaptation Were a Model? ShadowPEFT in 🤗 PEFT library
shadow-llm
• • 23Community Blog & Articles
Community Articles
view all
Your Agent Aced the Task. Will It Do It Again?
102
September 15, 2026 
llmrltrl
Async GRPO with LoRA across HF Jobs: a bucket, a proxy, and no NCCL
46
September 10, 2026 
gradioworkflowsautomatic1111
Rebuilding AUTOMATIC1111 with Gradio Workflow
62
September 10, 2026 
Safety for Whom? Refusing the Right Subset of a Topic, Not the Whole Topic
30
September 8, 2026 
NeoMME: an efficient Multimodal-native and Multilingual Encoder
110
September 3, 2026 
nlpguidecommunity
Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
129
September 3, 2026 
agentsopen-sourcedatasets
Give Your Coding Agents a Memory You Own
119
September 3, 2026 
rlgrpotrl
Training a coding model to paint watercolours with TRL and OpenEnv
70
September 3, 2026 
BenchMIRT: What are LLM benchmarks actually measuring?
24
September 1, 2026 
inferencejavascriptwebgpu
Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI
76
September 1, 2026 
audiospeechleaderboard
The Open ASR Leaderboard Adds Its First Global South Language
- +6
57
August 28, 2026 
nlpguidecommunity
Training and Finetuning Multi-Vector Embedding Models with Sentence Transformers
142
August 26, 2026 
Granite 4.2 LLMs: How They're Built
122
August 25, 2026 
Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original
68
August 25, 2026 NEW Articles from Team or Enterprise organizations will get promoted to the main section.
What If the Adaptation Were a Model? ShadowPEFT in 🤗 PEFT library
shadow-llm
• • 23适配器即模型 —— ShadowPEFT 现已集成到 🤗 PEFT 库
shadow-llm
• • 16Your Inference Server is Secretly a Learner: Reef Infrastructure for Continual Self-Improving Agents
quao627
• • 15OpenRouter Leaderboard: 425 Models by Price, Measured Speed and Korean Quality
ginigen-ai
• • 12IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license
ibm-research
• • 60Inside the JEV Ecosystem: 13 Answer Verifiers on One Test Set
mayafree
• • 11Your model already knows it's wrong. Asking costs 0.06 seconds and zero tokens.
FINAL-Bench
• • 11funes: Local Memory for Coding Agents, Built on Lance
ariG23498
• • 8Making open-source AI weather forecasting models easy to run
hugging-science
• • 37Optimum-Intel v2.2.0 & OpenVINO GenAI 2026.4.0: What's New
echarlaix
• • 7Uncensor any LLM with abliteration
mlabonne
• • 924BananaMind 2 Pro: We've (almost) matched SmolLM2 at 20x fewer tokens... Trained On a 5070 Ti
Banaxi-Tech
• • 6Open Yap 1K: 1,000 hours of full-duplex natural conversation, free for commercial use
TheAgenticDataCompany
• • 19KV Caching Explained: Optimizing Transformer Inference Efficiency
not-lain
• • 418Layer-Feedback Transformer (LFT)
Banaxi-Tech
• • 5One sandbox per rollout, or how labs run RL for agents in 2026
sergiopaniego
• • 10Per-tensor layout maps for GGUF quantization
bartowski
• • 37Deploy local agents everywhere with LFM2.5-2.6B
LiquidAI
• • 101DAVIDAU DELETED MY POST AND BANNED ME FOR ASKING ABOUT BENCHMARKS — HERE IS THE FULL ANALYSIS (9B + 27B + 390 MODELS)
DedeProGames
• • 23When Benchmark Numbers Become Marketing
DedeProGames
• • 4