Asilarkness
/

GLM-5.3-Flash-Uncensored-NVFP4-Pruning-62GB

Mixture of Experts

compressed-tensors

8-bit precision

Model card Files Files and versions

GLM-5.3-Flash-Uncensored-NVFP4-Pruning-62GB

Commit History

Fix short-conv layout: fuse q/k/v conv1d, rename forget_gate keys

7fcc98b
verified

Asilarkness commited on 28 days ago

Expert-pruned 288->80, 62GB, with model card

e2032a6
verified

Asilarkness commited on 28 days ago

initial commit

8027048
verified

Asilarkness commited on 28 days ago