Info

https://huggingface.co/stabilityai/stable-diffusion-3.5-large-turbo

Defaults for Turbo:

    num_inference_steps=4,
    guidance_scale=0.0,

Defaults for large

    num_inference_steps=28,
    guidance_scale=3.5,

Test 0 - Different seed variations

Prompt: photorealistic girl in bookshop choosing the book in romantic stories shelf. smiling

Execution: Time: 25m 47.53s | total 1651.35 pipeline 1535.92 preview 90.80 decode 11.56 prompt 7.30 offload 4.78 move 0.63 gc 0.30 | GPU 13394 MB 10% | RAM 18.12 GB 14%

CFG3.5, STEP 28	Seed: 1620085323	Seed: 1931701040	Seed: 2736029172	Seed: 4075624134
bookshop girl
hand and face
legs and shoes

Test 1 - Bookshop

Prompt: photorealistic girl in bookshop choosing the book in romantic stories shelf. smiling

AUG	8 (7 min)	16 (15 min)	20 (18 min)	32 (20 min)	50 (26 min)
CFG1
CFG2
CFG3
CFG4
CFG5
CFG6
CFG8

Test 2 - Face and hand

Prompt: Create a close-up photograph of a woman's face and hand, with her hand raised to her chin. She is wearing a white blazer and has a gold ring on her finger. Her nails are neatly manicured and her hair is pulled back into a low bun. She is smiling and has a radiant expression on her face. The background is a plain light gray color. The overall mood of the photo is elegant and sophisticated. The photo should have a soft, natural light and a slight warmth to it. The woman's hair is dark brown and pulled back into a low bun, with a few loose strands framing her face.

Time: 11m 51.44s | total 719.38 pipeline 706.57 callback 5.86 decode 4.84 prompt 1.78 gc 0.30 | GPU 45676 MB 36% | RAM 71.98 GB 57%

NOV	8	16	20	32 (11 minutes)	50
CFG1
CFG2
CFG3
CFG4
CFG6
CFG8

Test 3 - Legs

Prompt: Generate a photo of a woman's legs, with her feet crossed and wearing white high-heeled shoes with ribbons tied around her ankles. The shoes should have a pointed toe and a stiletto heel. The woman's legs should be smooth and tanned, with a slight sheen to them. The background should be a light gray color. The photo should be taken from a low angle, looking up at the woman's legs. The ribbons should be tied in a bow shape around the ankles. The shoes should have a red sole. The woman's legs should be slightly bent at the knee.

Time: 18m 35.12s | total 1124.50 pipeline 1110.25 callback 9.10 decode 4.83 gc 0.28 | GPU 45676 MB 36% | RAM 72.04 GB 57%

	8	16	20	32	50 (18 min)
CFG1
CFG2
CFG3
CFG4
CFG6
CFG8

Test 3.1 - With Negative Prompt

lowres, text, error, cropped, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, out of frame, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck, username, watermark, signature

Positive only

Negative added

Prompt: Generate a photo of a woman's legs, with her feet crossed and wearing white high-heeled shoes with ribbons tied around her ankles. The shoes should have a pointed toe and a stiletto heel. The woman's legs should be smooth and tanned, with a slight sheen to them. The background should be a light gray color. The photo should be taken from a low angle, looking up at the woman's legs. The ribbons should be tied in a bow shape around the ankles. The shoes should have a red sole. The woman's legs should be slightly bent at the knee.

Time: 18m 35.12s | total 1124.50 pipeline 1110.25 callback 9.10 decode 4.83 gc 0.28 | GPU 45676 MB 36% | RAM 72.04 GB 57%

Prompt: Generate a photo of a woman's legs, with her feet crossed and wearing white high-heeled shoes with ribbons tied around her ankles. The shoes should have a pointed toe and a stiletto heel. The woman's legs should be smooth and tanned, with a slight sheen to them. The background should be a light gray color. The photo should be taken from a low angle, looking up at the woman's legs. The ribbons should be tied in a bow shape around the ankles. The shoes should have a red sole. The woman's legs should be slightly bent at the knee.

Negative: lowres, text, error, cropped, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, out of frame, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck, username, watermark, signature

Time: 18m 33.60s | total 1125.09 pipeline 1108.65 callback 9.09 decode 4.90 prompt 2.03 gc 0.37 | GPU 45676 MB 36% | RAM 72.07 GB 58%

Test 4 - Other model covers

System info

Test Part 1

Mon Aug  4 19:48:58 2025
app: sdnext.git updated: 2025-08-01 hash: b56d508a url: https://github.com/vladmandic/sdnext.git/tree/master
arch: x86_64 cpu: x86_64 system: Linux release: 6.14.0-27-generic 
python: 3.12.3 Torch: 2.7.1+xpu
device: Intel(R) Arc(TM) Graphics (1) ipex: 
ram: free:122.51 used:2.82 total:125.33
xformers:  diffusers: 0.35.0.dev0 transformers: 4.54.1
active: xpu dtype: torch.bfloat16 vae: torch.bfloat16 unet: torch.bfloat16
base: Diffusers/stabilityai/stable-diffusion-3.5-large [ceddf0a7fd] refiner: none vae: none te: none unet: none
Backend: ipex Cross-attention: Scaled-Dot-Product

Test Part 2-3

Wed Nov 26 09:51:01 2025
app: sdnext.git updated: 2025-11-22 hash: ded5afc5d url: https://github.com/liutyi/sdnext/tree/pytorch
arch: x86_64 cpu: x86_64 system: Linux release: 6.14.0-36-generic
python: 3.12.3 Torch: 2.9.1+xpu
device: Intel(R) Arc(TM) Graphics (1) ipex: 
ram: free:120.52 used:4.81 total:125.33
xformers: diffusers: 0.36.0.dev0 transformers: 4.57.1
active: xpu dtype: torch.bfloat16 vae: torch.bfloat16 unet: torch.bfloat16
base: stabilityai/stable-diffusion-3.5-large refiner: none vae: none te: none unet: none
Backend: ipex Pipeline: native Cross-attention: Scaled-Dot-Product

Model

Model: Diffusers/stabilityai/stable-diffusion-3.5-large
Type: sd3
Class: StableDiffusion3Pipeline
Size: 0 bytes
Modified: 2025-08-04 11:51:22

Module Class Device DType Params Modules Config

vae	AutoencoderKL	cpu	torch.bfloat16	83819683	241	FrozenDict({'in_channels': 3, 'out_channels': 3, 'down_block_types': ['DownEncoderBlock2D', 'DownEncoderBlock2D', 'DownEncoderBlock2D', 'DownEncoderBlock2D'], 'up_block_types': ['UpDecoderBlock2D', 'UpDecoderBlock2D', 'UpDecoderBlock2D', 'UpDecoderBlock2D'], 'block_out_channels': [128, 256, 512, 512], 'layers_per_block': 2, 'act_fn': 'silu', 'latent_channels': 16, 'norm_num_groups': 32, 'sample_size': 1024, 'scaling_factor': 1.5305, 'shift_factor': 0.0609, 'latents_mean': None, 'latents_std': None, 'force_upcast': True, 'use_quant_conv': False, 'use_post_quant_conv': False, 'mid_block_add_attention': True, '_class_name': 'AutoencoderKL', '_diffusers_version': '0.31.0.dev0', '_name_or_path': '/mnt/models/Diffusers/models--stabilityai--stable-diffusion-3.5-large/snapshots/ceddf0a7fdf2064ea28e2213e3b84e4afa170a0f/vae'})
text_encoder	CLIPTextModelWithProjection	xpu:0	torch.bfloat16	123650304	153	CLIPTextConfig { "architectures": [ "CLIPTextModelWithProjection" ], "attention_dropout": 0.0, "bos_token_id": 0, "dropout": 0.0, "eos_token_id": 2, "hidden_act": "quick_gelu", "hidden_size": 768, "initializer_factor": 1.0, "initializer_range": 0.02, "intermediate_size": 3072, "layer_norm_eps": 1e-05, "max_position_embeddings": 77, "model_type": "clip_text_model", "num_attention_heads": 12, "num_hidden_layers": 12, "pad_token_id": 1, "projection_dim": 768, "torch_dtype": "bfloat16", "transformers_version": "4.54.1", "vocab_size": 49408 }
text_encoder_2	CLIPTextModelWithProjection	xpu:0	torch.bfloat16	694659840	393	CLIPTextConfig { "architectures": [ "CLIPTextModelWithProjection" ], "attention_dropout": 0.0, "bos_token_id": 0, "dropout": 0.0, "eos_token_id": 2, "hidden_act": "gelu", "hidden_size": 1280, "initializer_factor": 1.0, "initializer_range": 0.02, "intermediate_size": 5120, "layer_norm_eps": 1e-05, "max_position_embeddings": 77, "model_type": "clip_text_model", "num_attention_heads": 20, "num_hidden_layers": 32, "pad_token_id": 1, "projection_dim": 1280, "torch_dtype": "bfloat16", "transformers_version": "4.54.1", "vocab_size": 49408 }
text_encoder_3	T5EncoderModel	xpu:0	torch.bfloat16	4762310656	463	T5Config { "architectures": [ "T5EncoderModel" ], "classifier_dropout": 0.0, "d_ff": 10240, "d_kv": 64, "d_model": 4096, "decoder_start_token_id": 0, "dense_act_fn": "gelu_new", "dropout_rate": 0.1, "eos_token_id": 1, "feed_forward_proj": "gated-gelu", "initializer_factor": 1.0, "is_encoder_decoder": false, "is_gated_act": true, "layer_norm_epsilon": 1e-06, "model_type": "t5", "num_decoder_layers": 24, "num_heads": 64, "num_layers": 24, "output_past": true, "pad_token_id": 0, "relative_attention_max_distance": 128, "relative_attention_num_buckets": 32, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.54.1", "use_cache": false, "vocab_size": 32128 }
tokenizer	CLIPTokenizer	None	None	0	0	None
tokenizer_2	CLIPTokenizer	None	None	0	0	None
tokenizer_3	T5TokenizerFast	None	None	0	0	None
transformer	SD3Transformer2DModel	xpu:0	torch.bfloat16	8056627520	1456	FrozenDict({'sample_size': 128, 'patch_size': 2, 'in_channels': 16, 'num_layers': 38, 'attention_head_dim': 64, 'num_attention_heads': 38, 'joint_attention_dim': 4096, 'caption_projection_dim': 2432, 'pooled_projection_dim': 2048, 'out_channels': 16, 'pos_embed_max_size': 192, 'dual_attention_layers': (), 'qk_norm': 'rms_norm', '_use_default_values': ['dual_attention_layers'], '_class_name': 'SD3Transformer2DModel', '_diffusers_version': '0.31.0.dev0', '_name_or_path': '/mnt/models/Diffusers/models--stabilityai--stable-diffusion-3.5-large/snapshots/ceddf0a7fdf2064ea28e2213e3b84e4afa170a0f/transformer'})
scheduler	FlowMatchEulerDiscreteScheduler	None	None	0	0	FrozenDict({'num_train_timesteps': 1000, 'shift': 3.0, 'use_dynamic_shifting': False, 'base_shift': 0.5, 'max_shift': 1.15, 'base_image_seq_len': 256, 'max_image_seq_len': 4096, 'invert_sigmas': False, 'shift_terminal': None, 'use_karras_sigmas': False, 'use_exponential_sigmas': False, 'use_beta_sigmas': False, 'time_shift_type': 'exponential', 'stochastic_sampling': False, '_use_default_values': ['base_image_seq_len', 'invert_sigmas', 'use_exponential_sigmas', 'use_beta_sigmas', 'stochastic_sampling', 'max_image_seq_len', 'max_shift', 'time_shift_type', 'shift_terminal', 'use_karras_sigmas', 'base_shift', 'use_dynamic_shifting'], '_class_name': 'FlowMatchEulerDiscreteScheduler', '_diffusers_version': '0.29.0.dev0'})
image_encoder	NoneType	None	None	0	0	None
feature_extractor	NoneType	None	None	0	0	None
_name_or_path	str	None	None	0	0	None
_class_name	str	None	None	0	0	None
_diffusers_version	str	None	None	0	0	None

Page tree

Test 29 - Stable Diffusion 3.5 Large - steps and guidance