<h1 id="full-checkpoint-with-improved-te-do-not-load-additional-clipte-hxirwkac5">Full Checkpoint with improved TE do not load additional CLIP/TE</h1><h2 id="flux.1-(base-unet)-+-google-flan-z8h6fz4w1">FLUX.1 (Base UNET) + Google FLAN</h2><ul><li>All uploaded models sourced from 65GB Full FP32</li></ul><h1 id="-ugt6aovc9"></h1>Per the <a target="_blank" rel="ugc" href="https://www.apache.org/licenses/LICENSE-2.0">Apache 2.0 license</a> FLAN is attributed to Google

FLUX + FLAN (65GB Source)

fluxDevSchnellBaseUNET_fluxDevFLANFP16.safetensors

FLUX_Dev-FLAN-FP16

FLUX_DEV_FLAN_BF16

FLUX_Schnell_FLAN_BF16

FLUX_Schnell-FLAN_FP16

FLUX_Dev_FLAN_FP8

FLUX_Schnell_FLAN_FP8

FLUX_Dev-FLAN_NF4

FLUX_Schnell-FLAN_NF4

Full Checkpoint with improved TE do not load additional CLIP/TEFLUX.1 (Base UNET) + Google FLANThis model took the 42GB FP32 Google Flan T5xxl and quantized it with improved CLIP-L for Flux. To my knowledge no one else has posted or attempted this.<ul><li>Quantized from FP32 T5xxl (42GB 11B Parameter)</li><li>Base UNET no baked lora's or other changes</li><li>Full FP16 version is available.</li></ul>

⚡️FLUX Dev/Schnell (Base UNET) + Google FLAN FP16

Flux_Dev-FLAN-FP16.safetensors

Files with SHA256: b200a8a97d8ec9495abc0a42bd2254f248e8995d56210ff8c4ab974df96c154b