상위 문서로 이동 : AI Wiki

Training

Training Dataset

img_num	24	34	54	74	75
image
caption	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, wireless charger, microphone	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse	OTT, laptop, white tone, one monitor, mechanical keyboard, wireless mouse, trackpad, tablet	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, tablet	OTT, desktop, white tone, one monitor, mechanical keyboard, mouse

img_num	76	79	84	101	111
image
caption	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, tablet	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, wireless charger	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, tablet	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, wireless charger, microphone	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse

img_num	119	132	134	147	153
image
caption	OTT, laptop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, wireless charger, microphone	OTT, desktop, white tone, one monitor, mechanical keyboard, wireless mouse	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, tablet	OTT, laptop, white tone, one monitor, keyboard, track pad, wireless charger	OTT, laptop, white tone, one monitor, mechanical keyboard, wireless mouse, tablet

img_num	155	165	171	173	177
image
caption	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse	OTT, laptop, white tone, one monitor, speaker, mechanical keyboard	OTT, desktop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, wireless charger, tablet

Kohya_LoRA_Trainer_XL

• Base Model: stabilityai/stable-diffusion-xl-base-1.0
• VAE: Original VAE
• Network Category: LoRA_LierLa
• Network dim: 32
• Network alpha: 16
• Optimizer type: Prodigy
• Learning Rate: 0.5
• Epochs: 20
• Train Batch Size: 1
• Mixed Precision: fp16

Generating Test

• Base Model: sd_xl_base_1.0.safetensors
• VAE: sdxl_vae.safetensors

T4

• prompt: "OTT, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 25.28초
• 최대 VRAM 사용량: 10163.17 MB

• prompt: "OTT, laptop, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 25.43초
• 최대 VRAM 사용량: 10163.17 MB

• prompt: "OTT, laptop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 26.39초
• 최대 VRAM 사용량: 10163.17 MB

L4

• prompt: "OTT, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 12.55초
• 최대 VRAM 사용량: 10933.42 MB

• prompt: "OTT, laptop, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 11.70초
• 최대 VRAM 사용량: 10933.42 MB

• prompt: "OTT, laptop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 11.74초
• 최대 VRAM 사용량: 10933.42 MB

A100

• prompt: "OTT, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 4.93초
• 최대 VRAM 사용량: 10933.42 MB

• prompt: "OTT, laptop, white tone, one monitor, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 3.80초
• 최대 VRAM 사용량: 10933.42 MB

• prompt: "OTT, laptop, white tone, one monitor, speaker, mechanical keyboard, wireless mouse, highly detailed, 4k"
• negative_prompt: "blurry, low quality, deformed, cropped, watermark, text"
• num_inference_steps: 30
• guidance_scale: 7.5
• image size: 1024 x 1024
• 이미지 생성 소요 시간: 3.83초
• 최대 VRAM 사용량: 10933.42 MB