Info
https://huggingface.co/nvidia/Cosmos-Predict2-2B-Text2Image
executed on GMKTec EVO-T1 (Intel core ultra 9 285H) Intel Arc iGPU
Part 1 - Bookshop
Prompt: photorealistic girl in bookshop choosing the book in romantic stories shelf. smiling
Parameters: Steps: 20| Size: 1024x1024| Seed: 3286438823| CFG scale: 1.0| App: SD.Next| Version: 6ea881b| Pipeline: Cosmos2TextToImagePipeline| Operations: txt2img| Model: Cosmos-Predict2-2B-Text2Image
Time: 52.01s | total 55.67 pipeline 51.92 vae 1.98 te 1.16 callback 0.53 | GPU 18358 MB 15% | RAM 28.1 GB 23%
| STEPS: 8 | STEPS: 16 | STEPS: 20 | STEPS: 32 | STEPS: 50 | |
|---|---|---|---|---|---|
| CFG1 |
|
|
|
|
|
| CGF2 |
|
|
|
|
|
| CFG3 |
|
|
|
|
|
| CFG4 |
|
|
|
|
|
| CFG5 |
|
|
|
|
|
| CFG6 |
|
|
|
|
|
| CFG7 |
|
|
|
|
|
| CFG12 |
|
|
|
|
|
Part 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.
Parameters: Steps: 32| Size: 1024x1024| Seed: 3286438823| CFG scale: 1.0| App: SD.Next| Version: 6ea881b| Pipeline: Cosmos2TextToImagePipeline| Operations: txt2img| Model: Cosmos-Predict2-2B-Text2Image
285H Time: 1m 19.28s | total 81.38 pipeline 79.24 te 1.07 callback 0.85 | GPU 18360 MB 15% | RAM 27.86 GB 23%
| 8 | 16 | 32 | 64 | |
|---|---|---|---|---|
| CFG=1 |
|
|
|
|
| CFG=2 |
|
|
|
|
| CFG=6 |
|
|
|
|
| CFG=8 |
|
|
|
|
Part 3 - Legs and ribbon
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.
Parameters: Steps: 32| Size: 1024x1024| Seed: 3286438823| CFG scale: 2| App: SD.Next| Version: 6ea881b| Pipeline: Cosmos2TextToImagePipeline| Operations: txt2img| Model: Cosmos-Predict2-2B-Text2Image
Time: 2m 36.32s | total 159.50 pipeline 156.29 te 2.16 callback 0.83 | GPU 18360 MB 15% | RAM 27.91 GB 23%
| 8 | 16 | 32 | 64 | |
|---|---|---|---|---|
| CFG=1 |
|  |
|
|
| CFG=2 |
|
|
|
|
| CFG=4 |
|
|
|
|
| CFG=6 |
|
|
|
|
| CFG=8 |
|
|
|
|
Test 4 . Other model covers
Test 5 Art Prompts
Test 6. Seeds
System Info
Sun Dec 7 14:42:41 2025 app: sdnext.git updated: 2025-12-05 hash: 6ea881b10 url: https://github.com/liutyi/sdnext/tree/pytorch arch: x86_64 cpu: x86_64 system: Linux release: 6.17.0-7-generic python: 3.12.3 Torch 2.9.1+xpu device: Intel(R) Arc(TM) Graphics (1) ipex: ram: free:119.61 used:3.46 total:123.07 xformers: diffusers: 0.36.0.dev0 transformers: 4.57.1 active: xpu dtype: torch.bfloat16 vae: torch.bfloat16 unet: torch.bfloat16 base: nvidia/Cosmos-Predict2-2B-Text2Image refiner: none vae: none te: none unet: none Backend: ipex; Pipeline: native; Memory optimization: none; Cross-attention: Scaled-Dot-Product
Model Data
Model: nvidia/Cosmos-Predict2-2B-Text2Image
nvidia/Cosmos-Predict2-2B-Text2Image
| Module | Class | Device | Dtype | Quant | Params | Modules | Config |
|---|---|---|---|---|---|---|---|
| vae | AutoencoderKLWan | xpu:0 | torch.bfloat16 | None | 126892531 | 260 | FrozenDict({'base_dim': 96, 'decoder_base_dim': None, 'z_dim': 16, 'dim_mult': [1, 2, 4, 4], 'num_res_blocks': 2, 'attn_scales': [], 'temperal_downsample': [False, True, True], 'dropout': 0.0, 'latents_mean': [-0.7571, -0.7089, -0.9113, 0.1075, -0.1745, 0.9653, -0.1517, 1.5508, 0.4134, -0.0715, 0.5517, -0.3632, -0.1922, -0.9497, 0.2503, -0.2921], 'latents_std': [2.8184, 1.4541, 2.3275, 2.6558, 1.2196, 1.7708, 2.6052, 2.0743, 3.2687, 2.1526, 2.8652, 1.5579, 1.6382, 1.1253, 2.8251, 1.916], 'is_residual': False, 'in_channels': 3, 'out_channels': 3, 'patch_size': None, 'scale_factor_temporal': 4, 'scale_factor_spatial': 8, '_use_default_values': ['scale_factor_temporal', 'scale_factor_spatial', 'in_channels', 'patch_size', 'is_residual', 'decoder_base_dim', 'out_channels'], '_class_name': 'AutoencoderKLWan', '_diffusers_version': '0.34.0.dev0', '_name_or_path': '/mnt/models/Diffusers/models--nvidia--Cosmos-Predict2-2B-Text2Image/snapshots/acdb5fde992a73ef0355f287977d002cbfd127e0/vae'}) |
| text_encoder | T5EncoderModel | xpu:0 | torch.bfloat16 | None | 4864791552 | 439 | T5Config { "architectures": [ "T5EncoderModel" ], "classifier_dropout": 0.0, "d_ff": 65536, "d_kv": 128, "d_model": 1024, "decoder_start_token_id": 0, "dense_act_fn": "relu", "dropout_rate": 0.1, "dtype": "bfloat16", "eos_token_id": 1, "feed_forward_proj": "relu", "initializer_factor": 1.0, "is_encoder_decoder": false, "is_gated_act": false, "layer_norm_epsilon": 1e-06, "model_type": "t5", "n_positions": 512, "num_decoder_layers": 24, "num_heads": 128, "num_layers": 24, "output_past": true, "pad_token_id": 0, "relative_attention_max_distance": 128, "relative_attention_num_buckets": 32, "task_specific_params": { "summarization": { "early_stopping": true, "length_penalty": 2.0, "max_length": 200, "min_length": 30, "no_repeat_ngram_size": 3, "num_beams": 4, "prefix": "summarize: " }, "translation_en_to_de": { "early_stopping": true, "max_length": 300, "num_beams": 4, "prefix": "translate English to German: " }, "translation_en_to_fr": { "early_stopping": true, "max_length": 300, "num_beams": 4, "prefix": "translate English to French: " }, "translation_en_to_ro": { "early_stopping": true, "max_length": 300, "num_beams": 4, "prefix": "translate English to Romanian: " } }, "transformers_version": "4.57.1", "use_cache": false, "vocab_size": 32128 } |
| tokenizer | T5TokenizerFast | None | None | None | 0 | 0 | None |
| transformer | CosmosTransformer3DModel | xpu:0 | torch.bfloat16 | None | 1956405248 | 1138 | FrozenDict({'in_channels': 16, 'out_channels': 16, 'num_attention_heads': 16, 'attention_head_dim': 128, 'num_layers': 28, 'mlp_ratio': 4.0, 'text_embed_dim': 1024, 'adaln_lora_dim': 256, 'max_size': [128, 240, 240], 'patch_size': [1, 2, 2], 'rope_scale': [1.0, 4.0, 4.0], 'concat_padding_mask': True, 'extra_pos_embed_type': None, '_class_name': 'CosmosTransformer3DModel', '_diffusers_version': '0.34.0.dev0', '_name_or_path': 'nvidia/Cosmos-Predict2-2B-Text2Image'}) |
| scheduler | FlowMatchEulerDiscreteScheduler | None | None | None | 0 | 0 | FrozenDict({'num_train_timesteps': 1000, 'shift': 1.0, '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': True, 'use_exponential_sigmas': False, 'use_beta_sigmas': False, 'time_shift_type': 'exponential', 'stochastic_sampling': False, '_class_name': 'FlowMatchEulerDiscreteScheduler', '_diffusers_version': '0.34.0.dev0', 'final_sigmas_type': 'sigma_min', 'sigma_data': 1.0, 'sigma_max': 80.0, 'sigma_min': 0.002}) |
| safety_checker | Fake_safety_checker | None | None | None | 0 | 0 | None |
