Skip to content

clm:8b

2 TagsUpdated 8.2B params2048 contextEnglishApache-2.0by Contrastive-LM

CLM-v0.1-8B on Qwen3-8B, 0.357 on typed decisions; built for agent, game and tool-calling states. Needs a 20 GB GPU.

8.2b
ollaya run clm:8b --preset triage "I was charged twice for my subscription this month and want a refund."

Details

  • graph1f7b6781f500 · 7 MBonnx · clm · 8.2B · fp32
  • weights31d6a825ae35 · 4.0 GBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/model-00001-of-00005.safetensors
  • weights5991236cea6f · 4.0 GBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/model-00002-of-00005.safetensors
  • weightsc5185c4794be · 4.0 GBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/model-00003-of-00005.safetensors
  • weightsb5ee7de71fbf · 3.2 GBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/model-00004-of-00005.safetensors
  • weights20c2d6366ab8 · 1.2 GBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/model-00005-of-00005.safetensors
  • weightsb2b4a8c9c2d3 · 76 MBhuggingface.co/Contrastive-LM/CLM-v0.1-8B/resolve/e939398…/CLM_v0.1-8B.pt
  • tokenizeraeb13307a71a · 11 MBhuggingface.co/Qwen/Qwen3-8B/resolve/b968826…/tokenizer.json
  • decisiond2fc1f758f5d · 2 KB{"engine": "onnx", "family": "clm", "encoder": "", "layout": "clm-v1", …}
  • calibrationade0f04e3391 · 201 B{"temperature": [1.0, 1.0, 1.0]}
  • licensef5bb8e0f6419 · 10 KBCLM-v0.1-8B by Contrastive-LM (https://huggingface.co/Contrastive-LM/CLM-v0.1-8B), state and action projection heads: Apache-2.0.

Every layer is checked against its sha256 when it is pulled. Weights and tokenizers download from the model author's Hugging Face repository at a pinned commit; Ollaya never re-hosts them.

CLM readme and all models