SD

VXPv2 (new model Type-R released)

VXPv2 (new model Type-R released) - AI Model cover image
Dimuat naik pada
14 Dis 2023, 1:57 PTG
Kegunaan
80.5k
Ulasan
Cemerlang  (12)
Langkah Pensampelan

20

Kaedah Pensampelan

DPM++ 2M Karras

Skala CFG

8

Negatif

worst quality, large head, low quality, extra digits, bad eye, bad anatomy, EasyNegativeV2, ng_deepnegative_v1_75t,(extra fingers:1.5), (deformed hands:1.5), (polydactyl:1.5)

Kebenaran
  • Allow image generation & sharing
  • Benarkan pengguna memuat turun model anda
  • Kegunaan komersial

Keterangan

This model generate high detailed images without LoRAs like "more detail" or "add detail". Be careful to use them or you obtain a over detailed images, my suggestion is to start without them, if needed use them with 0.1-0.3 range. Have fun😘

V1 and V2 are two different models that I'm developing in parallel

V1 have anime stile, V2 have more realistic style.

the purpose of this model is to make high-quality images by reducing the use of LoRAs and simplifying prompts; in fact, prompts such as "masterpiece", "top qualiti", etc. are no longer needed...😉 V1 : semi realistic anime style model. V1.1 : Improvemento ov V1, Better hand finger, more detail on Eyes, general color improvement. V2 : graphics rendering has been further improved, also improved the details of hands and fingers. V2.1 : evolution of V2, slightly improved general image quality. The face now less childish. Slightly improved eyes quality. V2.2 : Evolution of V2.1, slightly improved general image quality. improved hair quality. V2.3 : Enhancement of V2.2, changed color rendering, now more intense, improved Eye detail. V2.31 : Correction of V2.3, 2.3 will be deprecated, please use 2.31 V2.4 : Deprecated V2.41 : Enhanced light, enhanced backgrounds, calibrated light and shadow. General image quality improved V2.45 : Evolution of 2.41, internal VAE optimized, reduced "black spot", general improvement, character model, background ect...

Komen8