Model

Pruned safetensors weights for KayLazy's LazyMix+ v4.0 SD1.5 checkpoint.
Aim is to save storage by cutting down from over 5 GB down to about 2 GB with minimal inference quality loss.

Details

Produced from the full fp32 checkpoint file, using lopho/stable-diffusion-prune.
As a sanity check, pruning the published fp16 weights as input also produces the same exact file with --fp16 flag, regardless of --ema flag.

  • Inputs
    • Full Model fp32 lazymixRealAmateur_v40.safetensors
      SHA256: 352d682c69c8280e3a4fb73d3129f5ab33d0a65c43b15dd75cc25bd4b3f0806d
      CivArchive
    • Full Model fp16 lazymixRealAmateur_v40.safetensors
      SHA256: d5fd15bf727be286564ed6af5c9a612348f328eadd9067784cd57fcb34891546
      CivArchive
  • Outputs
    • Pruned Model fp32 lazymixRealAmateur_v4.0_pruned_fp32.safetensors
      SHA256: ee211a9d09e710a509116460103909895fdc40d62cb62aede376aa13cf15976a
      CivArchive
    • Pruned Model fp16 lazymixRealAmateur_v4.0_pruned_fp16.safetensors
      SHA256: f448b74660927ad06d2aa87e789562674350d6f38e7f7010846e0e0437841074
      CivArchive

Legal

This is a derivative of a permissively-licenced work.
Original publisher's CreativeML Open RAIL-M license inherited.
Original publisher's KayLazy authorship uncontested.
lopho/stable-diffusion-prune licenced under AGPL-3.0 not derivated upon.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Erquint/lazymixRealAmateur_v40_pruned_fp16

Base model

minaiosu/kaylazy
Finetuned
(1)
this model