CoolFace
Modelpublic

Chaman1234/Sparse-AST-BWM

sourceHugging Faceapache-2.0updated 6d agoView on Hugging Face
0likes1.7kdownloads
28 commits on main
7c1001d6d ago

Add verified Family Old/New step 51375 SafeTensors checkpoint

Chaman1234
c4f74dd8d ago

Document step 46375 deep OOD results and token budget guidance

Chaman1234
6376b358d ago

Replace Family Old/New checkpoint with verified step 46375 safetensors

Chaman1234
e20f0a29d ago

Add Family Old/New Colab loader sources

Chaman1234
d8073ee9d ago

Add Family Old/New Colab loader sources

Chaman1234
9c2417b9d ago

Replace family-oldnew checkpoint with step 32375 safetensors

Chaman1234
b7987b310d ago

Document constrained graph harness evaluation

Chaman1234
6925e4911d ago

Upload family-oldnew-2m/checkpoint_family_oldnew_2m_snapshot_step_14550.safetensors with huggingface_hub

Chaman1234
58a4f2d12d ago

Document safe oldnew connections with other variants

Chaman1234
ce0347712d ago

Clarify family-oldnew experiment and remove MoE claims

Chaman1234
8c0843812d ago

Correct model card metadata and architecture documentation

Chaman1234
799f5dc12d ago

Fix model card YAML metadata front matter

Chaman1234
3f9b14b12d ago

Add reproducible BlenderBench proxy evaluator

Chaman1234
82bbc4012d ago

Add BlenderBench evaluation report

Chaman1234
248a1e512d ago

Add BlenderBench 27-task proxy evaluation results

Chaman1234
a74c60912d ago

Clarify Top-3 K=N is dense ensemble control and sparse alternatives

Chaman1234
a4f540812d ago

Add three-expert Top-3 Blender prompt evaluation results

Chaman1234
9d6f97612d ago

Add corrected three-expert Top-3 evaluator

Chaman1234
45f520112d ago

Document corrected Top-3 100M-family 100M 200M Blender MoE results

Chaman1234
974dbd912d ago

Document Sparse-AST BWM architecture, BMW math, and family old-new MoE

Chaman1234
805d92712d ago

Add reproducible real Blender prompt MoE evaluator

Chaman1234
719594612d ago

Add real Blender prompt MoE evaluation report

Chaman1234
0326acc12d ago

Add corrected family old-new safetensors 1M token snapshot

Chaman1234
4ece67414d ago

Update README: 200M completion, Hard Blender 5.1 Evals, and 3D generation pipeline

Chaman1234
08e73b415d ago

Update README with Qwen2.5-Coder 3B local Ollama hybrid architecture guide and benchmarks

Chaman1234
05912a915d ago

Add Hybrid Ollama and Pretrained LLM integration guide

Chaman1234
fec715815d ago

Publish unified Sparse-AST / BWM model family under one repository

Chaman1234
8b39bb515d ago

initial commit

Chaman1234