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
      • Learn
      • Discord
      • Forum
      • GitHub
    • Solutions
      • Team & Enterprise
      • Hugging Face PRO
      • Enterprise Support
      • Inference Providers
      • Inference Endpoints
      • Storage Buckets

  • Log In
  • Sign Up

GenomaLabs-com
/
kv-cache-eviction-mla

Transformers
kv-cache
h2o
mla
long-context
inference-optimization
deepseek
kimi
Model card Files Files and versions
xet
Community

Instructions to use GenomaLabs-com/kv-cache-eviction-mla with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

  • Libraries
  • Transformers

    How to use GenomaLabs-com/kv-cache-eviction-mla with Transformers:

    # Load model directly
    from transformers import AutoModel
    model = AutoModel.from_pretrained("GenomaLabs-com/kv-cache-eviction-mla", dtype="auto")
  • Notebooks
  • Google Colab
  • Kaggle
kv-cache-eviction-mla / scripts
55.2 kB
Ctrl+K
Ctrl+K
  • 4 contributors
History: 6 commits
GENOMA LABS / research
Round 3b prep: control-validation driver (plain HF transformers, no AirLLM)
754890f 10 days ago
  • kimi_layer_eviction_demo.py
    8.52 kB
    B3a-pivot: real Kimi K2.6 weights demo - eviction policy validated on actual MLA attention distribution 10 days ago
  • niah_harness.py
    14.6 kB
    Round 3a: RULER NIAH single-needle harness (logic validated) 10 days ago
  • run_control_niah.py
    5.42 kB
    Round 3b prep: control-validation driver (plain HF transformers, no AirLLM) 10 days ago
  • run_kimi_ruler_sweep.py
    10.7 kB
    Round 4 driver: AirLLM Kimi K2.6 + install_kv_eviction + NIAH sweep 10 days ago
  • validate_eviction_random_init.py
    8.1 kB
    B1 validation: multi-step eviction test + transformers compatibility note 10 days ago
  • validate_install_on_real_kimi_layer.py
    7.9 kB
    Day 1 of 7-day Kimi K2.6 + RULER 128K sprint: transformers 5.x port + real-weights install validation 10 days ago