Hugging Face's logo Hugging Face
  • Models
  • Datasets
  • Spaces
  • Buckets new
  • Docs
  • Enterprise
  • Pricing
    • Website
      • Tasks
      • HuggingChat
      • Collections
      • Languages
      • Organizations
    • Community
      • Blog
      • Posts
      • Daily Papers
      • Hardware
      • Learn
      • Discord
      • Forum
      • GitHub
    • Solutions
      • Team & Enterprise
      • Hugging Face PRO
      • Enterprise Support
      • Inference Providers
      • Inference Endpoints
      • Storage Buckets

  • Log In
  • Sign Up
kshitijthakkar 's Collections
ICML 2026 Reproductions - agent-repro challenge
Kirigami: Zero-Shot Expert Carves of Qwen3.6
DeepSeek V4 Replicas
mcp-server-bench
Qwen3.5 Dense-to-MoE Weight Transfer
Large MoE Architecture Search (1B-2B)
Mobile MoE Architecture Search
OutageOdyssey
TraceMind-AI
Loggenix-MOE

Kirigami: Zero-Shot Expert Carves of Qwen3.6

updated 15 days ago

Qwen3.6-35B-A3B carved by expert importance to fit one consumer GPU. No training, no calibration. 800 tok/s on a laptop 5090.

Upvote
-

  • kshitijthakkar/Kirigami-Qwen3.6-20B-A3B-NVFP4

    Text Generation • 14B • Updated 19 days ago • 70 • 1

  • kshitijthakkar/Kirigami-Qwen3.6-24B-A3B-NVFP4

    Text Generation • 16B • Updated 19 days ago • 284

  • kshitijthakkar/Kirigami-Qwen3.6-28B-A3B-NVFP4

    Text Generation • 19B • Updated 19 days ago • 21 • 1

  • Running

    Kirigami Journey

    🪷

    How we carved a 35B MoE to fit a 24GB GPU — zero training

Upvote
-
  • Collection guide
  • Browse collections
Company
TOS Privacy About Careers
Website
Models Datasets Spaces Pricing Docs