diff --git a/tests/lora/test_lora_layers_sd.py b/tests/lora/test_lora_layers_sd.py index 0f606a056f67..5d79bb0c50bc 100644 --- a/tests/lora/test_lora_layers_sd.py +++ b/tests/lora/test_lora_layers_sd.py @@ -20,7 +20,6 @@ import torch import torch.nn as nn from huggingface_hub import hf_hub_download -from huggingface_hub.repocard import RepoCard from safetensors.torch import load_file from transformers import CLIPTextModel, CLIPTokenizer @@ -103,7 +102,7 @@ def tearDown(self): @slow @require_torch_gpu def test_integration_move_lora_cpu(self): - path = "runwayml/stable-diffusion-v1-5" + path = "Jiali/stable-diffusion-1.5" lora_id = "takuma104/lora-test-text-encoder-lora-target" pipe = StableDiffusionPipeline.from_pretrained(path, torch_dtype=torch.float16) @@ -162,7 +161,7 @@ def test_integration_move_lora_cpu(self): def test_integration_move_lora_dora_cpu(self): from peft import LoraConfig - path = "Lykon/dreamshaper-8" + path = "Jiali/stable-diffusion-1.5" unet_lora_config = LoraConfig( init_lora_weights="gaussian", target_modules=["to_k", "to_q", "to_v", "to_out.0"], @@ -222,7 +221,7 @@ def tearDown(self): torch.cuda.empty_cache() def test_integration_logits_with_scale(self): - path = "runwayml/stable-diffusion-v1-5" + path = "Jiali/stable-diffusion-1.5" lora_id = "takuma104/lora-test-text-encoder-lora-target" pipe = StableDiffusionPipeline.from_pretrained(path, torch_dtype=torch.float32) @@ -254,7 +253,7 @@ def test_integration_logits_with_scale(self): release_memory(pipe) def test_integration_logits_no_scale(self): - path = "runwayml/stable-diffusion-v1-5" + path = "Jiali/stable-diffusion-1.5" lora_id = "takuma104/lora-test-text-encoder-lora-target" pipe = StableDiffusionPipeline.from_pretrained(path, torch_dtype=torch.float32) @@ -284,8 +283,8 @@ def test_dreambooth_old_format(self): generator = torch.Generator("cpu").manual_seed(0) lora_model_id = "hf-internal-testing/lora_dreambooth_dog_example" - card = RepoCard.load(lora_model_id) - base_model_id = card.data.to_dict()["base_model"] + + base_model_id = "Jiali/stable-diffusion-1.5" pipe = StableDiffusionPipeline.from_pretrained(base_model_id, safety_checker=None) pipe = pipe.to(torch_device) @@ -308,8 +307,8 @@ def test_dreambooth_text_encoder_new_format(self): generator = torch.Generator().manual_seed(0) lora_model_id = "hf-internal-testing/lora-trained" - card = RepoCard.load(lora_model_id) - base_model_id = card.data.to_dict()["base_model"] + + base_model_id = "Jiali/stable-diffusion-1.5" pipe = StableDiffusionPipeline.from_pretrained(base_model_id, safety_checker=None) pipe = pipe.to(torch_device) @@ -420,7 +419,7 @@ def test_a1111_with_sequential_cpu_offload(self): def test_kohya_sd_v15_with_higher_dimensions(self): generator = torch.Generator().manual_seed(0) - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None).to( + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None).to( torch_device ) lora_model_id = "hf-internal-testing/urushisato-lora" @@ -444,8 +443,8 @@ def test_vanilla_funetuning(self): generator = torch.Generator().manual_seed(0) lora_model_id = "hf-internal-testing/sd-model-finetuned-lora-t4" - card = RepoCard.load(lora_model_id) - base_model_id = card.data.to_dict()["base_model"] + + base_model_id = "Jiali/stable-diffusion-1.5" pipe = StableDiffusionPipeline.from_pretrained(base_model_id, safety_checker=None) pipe = pipe.to(torch_device) @@ -468,7 +467,7 @@ def test_unload_kohya_lora(self): prompt = "masterpiece, best quality, mountain" num_inference_steps = 2 - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None).to( + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None).to( torch_device ) initial_images = pipe( @@ -506,7 +505,7 @@ def test_load_unload_load_kohya_lora(self): prompt = "masterpiece, best quality, mountain" num_inference_steps = 2 - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None).to( + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None).to( torch_device ) initial_images = pipe( @@ -548,9 +547,9 @@ def test_load_unload_load_kohya_lora(self): def test_not_empty_state_dict(self): # Makes sure https://github.com/huggingface/diffusers/issues/7054 does not happen again - pipe = AutoPipelineForText2Image.from_pretrained( - "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16 - ).to(torch_device) + pipe = AutoPipelineForText2Image.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16).to( + torch_device + ) pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config) cached_file = hf_hub_download("hf-internal-testing/lcm-lora-test-sd-v1-5", "test_lora.safetensors") @@ -562,9 +561,9 @@ def test_not_empty_state_dict(self): def test_load_unload_load_state_dict(self): # Makes sure https://github.com/huggingface/diffusers/issues/7054 does not happen again - pipe = AutoPipelineForText2Image.from_pretrained( - "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16 - ).to(torch_device) + pipe = AutoPipelineForText2Image.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16).to( + torch_device + ) pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config) cached_file = hf_hub_download("hf-internal-testing/lcm-lora-test-sd-v1-5", "test_lora.safetensors") @@ -581,7 +580,7 @@ def test_load_unload_load_state_dict(self): release_memory(pipe) def test_sdv1_5_lcm_lora(self): - pipe = DiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16) + pipe = DiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16) pipe.to(torch_device) pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config) @@ -609,7 +608,7 @@ def test_sdv1_5_lcm_lora(self): release_memory(pipe) def test_sdv1_5_lcm_lora_img2img(self): - pipe = AutoPipelineForImage2Image.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16) + pipe = AutoPipelineForImage2Image.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16) pipe.to(torch_device) pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config) @@ -650,7 +649,7 @@ def test_sd_load_civitai_empty_network_alpha(self): This test simply checks that loading a LoRA with an empty network alpha works fine See: https://github.com/huggingface/diffusers/issues/5606 """ - pipeline = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + pipeline = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5") pipeline.enable_sequential_cpu_offload() civitai_path = hf_hub_download("ybelkada/test-ahi-civitai", "ahi_lora_weights.safetensors") pipeline.load_lora_weights(civitai_path, adapter_name="ahri") diff --git a/tests/models/autoencoders/test_models_vae.py b/tests/models/autoencoders/test_models_vae.py index 1b1c9b35214e..13e9bb1ba569 100644 --- a/tests/models/autoencoders/test_models_vae.py +++ b/tests/models/autoencoders/test_models_vae.py @@ -1051,7 +1051,7 @@ def test_encode_decode(self): def test_sd(self): vae = ConsistencyDecoderVAE.from_pretrained("openai/consistency-decoder") # TODO - update - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", vae=vae, safety_checker=None) + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", vae=vae, safety_checker=None) pipe.to(torch_device) out = pipe( @@ -1099,7 +1099,7 @@ def test_sd_f16(self): "openai/consistency-decoder", torch_dtype=torch.float16 ) # TODO - update pipe = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", torch_dtype=torch.float16, vae=vae, safety_checker=None, @@ -1124,7 +1124,7 @@ def test_sd_f16(self): def test_vae_tiling(self): vae = ConsistencyDecoderVAE.from_pretrained("openai/consistency-decoder", torch_dtype=torch.float16) pipe = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", vae=vae, safety_checker=None, torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", vae=vae, safety_checker=None, torch_dtype=torch.float16 ) pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) diff --git a/tests/models/unets/test_models_unet_2d_condition.py b/tests/models/unets/test_models_unet_2d_condition.py index df88e7960ba9..f91686925024 100644 --- a/tests/models/unets/test_models_unet_2d_condition.py +++ b/tests/models/unets/test_models_unet_2d_condition.py @@ -1376,7 +1376,7 @@ def test_compvis_sd_v1_4_fp16(self, seed, timestep, expected_slice): @require_torch_accelerator @skip_mps def test_compvis_sd_v1_5(self, seed, timestep, expected_slice): - model = self.get_unet_model(model_id="runwayml/stable-diffusion-v1-5") + model = self.get_unet_model(model_id="Jiali/stable-diffusion-1.5") latents = self.get_latents(seed) encoder_hidden_states = self.get_encoder_hidden_states(seed) @@ -1404,7 +1404,7 @@ def test_compvis_sd_v1_5(self, seed, timestep, expected_slice): ) @require_torch_accelerator_with_fp16 def test_compvis_sd_v1_5_fp16(self, seed, timestep, expected_slice): - model = self.get_unet_model(model_id="runwayml/stable-diffusion-v1-5", fp16=True) + model = self.get_unet_model(model_id="Jiali/stable-diffusion-1.5", fp16=True) latents = self.get_latents(seed, fp16=True) encoder_hidden_states = self.get_encoder_hidden_states(seed, fp16=True) @@ -1433,7 +1433,7 @@ def test_compvis_sd_v1_5_fp16(self, seed, timestep, expected_slice): @require_torch_accelerator @skip_mps def test_compvis_sd_inpaint(self, seed, timestep, expected_slice): - model = self.get_unet_model(model_id="runwayml/stable-diffusion-inpainting") + model = self.get_unet_model(model_id="botp/stable-diffusion-v1-5-inpainting") latents = self.get_latents(seed, shape=(4, 9, 64, 64)) encoder_hidden_states = self.get_encoder_hidden_states(seed) @@ -1461,7 +1461,7 @@ def test_compvis_sd_inpaint(self, seed, timestep, expected_slice): ) @require_torch_accelerator_with_fp16 def test_compvis_sd_inpaint_fp16(self, seed, timestep, expected_slice): - model = self.get_unet_model(model_id="runwayml/stable-diffusion-inpainting", fp16=True) + model = self.get_unet_model(model_id="botp/stable-diffusion-v1-5-inpainting", fp16=True) latents = self.get_latents(seed, shape=(4, 9, 64, 64), fp16=True) encoder_hidden_states = self.get_encoder_hidden_states(seed, fp16=True) diff --git a/tests/pipelines/amused/test_amused.py b/tests/pipelines/amused/test_amused.py index 9a9e2551d642..32f3e13ad911 100644 --- a/tests/pipelines/amused/test_amused.py +++ b/tests/pipelines/amused/test_amused.py @@ -13,7 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. - import unittest import numpy as np @@ -21,7 +20,12 @@ from transformers import CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer from diffusers import AmusedPipeline, AmusedScheduler, UVit2DModel, VQModel -from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, slow, torch_device +from diffusers.utils.testing_utils import ( + enable_full_determinism, + require_torch_gpu, + slow, + torch_device, +) from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_PARAMS from ..test_pipelines_common import PipelineTesterMixin @@ -65,9 +69,7 @@ def get_dummy_components(self): vqvae = VQModel( act_fn="silu", block_out_channels=[8], - down_block_types=[ - "DownEncoderBlock2D", - ], + down_block_types=["DownEncoderBlock2D"], in_channels=3, latent_channels=8, layers_per_block=1, @@ -75,9 +77,7 @@ def get_dummy_components(self): num_vq_embeddings=8, out_channels=3, sample_size=8, - up_block_types=[ - "UpDecoderBlock2D", - ], + up_block_types=["UpDecoderBlock2D"], mid_block_add_attention=False, lookup_from_codebook=True, ) @@ -96,7 +96,6 @@ def get_dummy_components(self): ) text_encoder = CLIPTextModelWithProjection(text_encoder_config) tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip") - components = { "transformer": transformer, "scheduler": scheduler, @@ -135,47 +134,37 @@ class AmusedPipelineSlowTests(unittest.TestCase): def test_amused_256(self): pipe = AmusedPipeline.from_pretrained("amused/amused-256") pipe.to(torch_device) - image = pipe("dog", generator=torch.Generator().manual_seed(0), num_inference_steps=2, output_type="np").images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) - expected_slice = np.array([0.4011, 0.3992, 0.3790, 0.3856, 0.3772, 0.3711, 0.3919, 0.3850, 0.3625]) - assert np.abs(image_slice - expected_slice).max() < 3e-3 + expected_slice = np.array([0.4011, 0.3992, 0.379, 0.3856, 0.3772, 0.3711, 0.3919, 0.385, 0.3625]) + assert np.abs(image_slice - expected_slice).max() < 0.003 def test_amused_256_fp16(self): pipe = AmusedPipeline.from_pretrained("amused/amused-256", variant="fp16", torch_dtype=torch.float16) pipe.to(torch_device) - image = pipe("dog", generator=torch.Generator().manual_seed(0), num_inference_steps=2, output_type="np").images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) expected_slice = np.array([0.0554, 0.05129, 0.0344, 0.0452, 0.0476, 0.0271, 0.0495, 0.0527, 0.0158]) - assert np.abs(image_slice - expected_slice).max() < 7e-3 + assert np.abs(image_slice - expected_slice).max() < 0.007 def test_amused_512(self): pipe = AmusedPipeline.from_pretrained("amused/amused-512") pipe.to(torch_device) - image = pipe("dog", generator=torch.Generator().manual_seed(0), num_inference_steps=2, output_type="np").images - image_slice = image[0, -3:, -3:, -1].flatten() assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.9960, 0.9960, 0.9946, 0.9980, 0.9947, 0.9932, 0.9960, 0.9961, 0.9947]) - assert np.abs(image_slice - expected_slice).max() < 3e-3 + expected_slice = np.array([0.1199, 0.1171, 0.1229, 0.1188, 0.1210, 0.1147, 0.1260, 0.1346, 0.1152]) + assert np.abs(image_slice - expected_slice).max() < 0.003 def test_amused_512_fp16(self): pipe = AmusedPipeline.from_pretrained("amused/amused-512", variant="fp16", torch_dtype=torch.float16) pipe.to(torch_device) - image = pipe("dog", generator=torch.Generator().manual_seed(0), num_inference_steps=2, output_type="np").images - image_slice = image[0, -3:, -3:, -1].flatten() assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.9983, 1.0, 1.0, 1.0, 1.0, 0.9989, 0.9994, 0.9976, 0.9977]) - assert np.abs(image_slice - expected_slice).max() < 3e-3 + expected_slice = np.array([0.1509, 0.1492, 0.1531, 0.1485, 0.1501, 0.1465, 0.1581, 0.1690, 0.1499]) + assert np.abs(image_slice - expected_slice).max() < 0.003 diff --git a/tests/pipelines/amused/test_amused_img2img.py b/tests/pipelines/amused/test_amused_img2img.py index 24bc34d330e9..c647a5aa304e 100644 --- a/tests/pipelines/amused/test_amused_img2img.py +++ b/tests/pipelines/amused/test_amused_img2img.py @@ -13,7 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. - import unittest import numpy as np @@ -22,7 +21,12 @@ from diffusers import AmusedImg2ImgPipeline, AmusedScheduler, UVit2DModel, VQModel from diffusers.utils import load_image -from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, slow, torch_device +from diffusers.utils.testing_utils import ( + enable_full_determinism, + require_torch_gpu, + slow, + torch_device, +) from ..pipeline_params import TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS, TEXT_GUIDED_IMAGE_VARIATION_PARAMS from ..test_pipelines_common import PipelineTesterMixin @@ -35,9 +39,7 @@ class AmusedImg2ImgPipelineFastTests(PipelineTesterMixin, unittest.TestCase): pipeline_class = AmusedImg2ImgPipeline params = TEXT_GUIDED_IMAGE_VARIATION_PARAMS - {"height", "width", "latents"} batch_params = TEXT_GUIDED_IMAGE_VARIATION_BATCH_PARAMS - required_optional_params = PipelineTesterMixin.required_optional_params - { - "latents", - } + required_optional_params = PipelineTesterMixin.required_optional_params - {"latents"} def get_dummy_components(self): torch.manual_seed(0) @@ -69,19 +71,15 @@ def get_dummy_components(self): vqvae = VQModel( act_fn="silu", block_out_channels=[8], - down_block_types=[ - "DownEncoderBlock2D", - ], + down_block_types=["DownEncoderBlock2D"], in_channels=3, latent_channels=8, layers_per_block=1, norm_num_groups=8, - num_vq_embeddings=32, # reducing this to 16 or 8 -> RuntimeError: "cdist_cuda" not implemented for 'Half' + num_vq_embeddings=32, out_channels=3, sample_size=8, - up_block_types=[ - "UpDecoderBlock2D", - ], + up_block_types=["UpDecoderBlock2D"], mid_block_add_attention=False, lookup_from_codebook=True, ) @@ -100,7 +98,6 @@ def get_dummy_components(self): ) text_encoder = CLIPTextModelWithProjection(text_encoder_config) tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip") - components = { "transformer": transformer, "scheduler": scheduler, @@ -139,13 +136,11 @@ class AmusedImg2ImgPipelineSlowTests(unittest.TestCase): def test_amused_256(self): pipe = AmusedImg2ImgPipeline.from_pretrained("amused/amused-256") pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains.jpg") .resize((256, 256)) .convert("RGB") ) - image = pipe( "winter mountains", image, @@ -153,24 +148,19 @@ def test_amused_256(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) - expected_slice = np.array([0.9993, 1.0, 0.9996, 1.0, 0.9995, 0.9925, 0.9990, 0.9954, 1.0]) - - assert np.abs(image_slice - expected_slice).max() < 1e-2 + expected_slice = np.array([0.9993, 1.0, 0.9996, 1.0, 0.9995, 0.9925, 0.999, 0.9954, 1.0]) + assert np.abs(image_slice - expected_slice).max() < 0.01 def test_amused_256_fp16(self): pipe = AmusedImg2ImgPipeline.from_pretrained("amused/amused-256", torch_dtype=torch.float16, variant="fp16") pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains.jpg") .resize((256, 256)) .convert("RGB") ) - image = pipe( "winter mountains", image, @@ -178,24 +168,19 @@ def test_amused_256_fp16(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) - expected_slice = np.array([0.9980, 0.9980, 0.9940, 0.9944, 0.9960, 0.9908, 1.0, 1.0, 0.9986]) - - assert np.abs(image_slice - expected_slice).max() < 1e-2 + expected_slice = np.array([0.998, 0.998, 0.994, 0.9944, 0.996, 0.9908, 1.0, 1.0, 0.9986]) + assert np.abs(image_slice - expected_slice).max() < 0.01 def test_amused_512(self): pipe = AmusedImg2ImgPipeline.from_pretrained("amused/amused-512") pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains.jpg") .resize((512, 512)) .convert("RGB") ) - image = pipe( "winter mountains", image, @@ -203,23 +188,20 @@ def test_amused_512(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.1344, 0.0985, 0.0, 0.1194, 0.1809, 0.0765, 0.0854, 0.1371, 0.0933]) + expected_slice = np.array([0.2809, 0.1879, 0.2027, 0.2418, 0.1852, 0.2145, 0.2484, 0.2425, 0.2317]) assert np.abs(image_slice - expected_slice).max() < 0.1 def test_amused_512_fp16(self): pipe = AmusedImg2ImgPipeline.from_pretrained("amused/amused-512", variant="fp16", torch_dtype=torch.float16) pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains.jpg") .resize((512, 512)) .convert("RGB") ) - image = pipe( "winter mountains", image, @@ -227,9 +209,8 @@ def test_amused_512_fp16(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.1536, 0.1767, 0.0227, 0.1079, 0.2400, 0.1427, 0.1511, 0.1564, 0.1542]) + expected_slice = np.array([0.2795, 0.1867, 0.2028, 0.2450, 0.1856, 0.2140, 0.2473, 0.2406, 0.2313]) assert np.abs(image_slice - expected_slice).max() < 0.1 diff --git a/tests/pipelines/amused/test_amused_inpaint.py b/tests/pipelines/amused/test_amused_inpaint.py index d0c1ed09c706..4a8d501450bb 100644 --- a/tests/pipelines/amused/test_amused_inpaint.py +++ b/tests/pipelines/amused/test_amused_inpaint.py @@ -13,7 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. - import unittest import numpy as np @@ -22,7 +21,12 @@ from diffusers import AmusedInpaintPipeline, AmusedScheduler, UVit2DModel, VQModel from diffusers.utils import load_image -from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu, slow, torch_device +from diffusers.utils.testing_utils import ( + enable_full_determinism, + require_torch_gpu, + slow, + torch_device, +) from ..pipeline_params import TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS, TEXT_GUIDED_IMAGE_INPAINTING_PARAMS from ..test_pipelines_common import PipelineTesterMixin @@ -35,9 +39,7 @@ class AmusedInpaintPipelineFastTests(PipelineTesterMixin, unittest.TestCase): pipeline_class = AmusedInpaintPipeline params = TEXT_GUIDED_IMAGE_INPAINTING_PARAMS - {"width", "height"} batch_params = TEXT_GUIDED_IMAGE_INPAINTING_BATCH_PARAMS - required_optional_params = PipelineTesterMixin.required_optional_params - { - "latents", - } + required_optional_params = PipelineTesterMixin.required_optional_params - {"latents"} def get_dummy_components(self): torch.manual_seed(0) @@ -50,7 +52,7 @@ def get_dummy_components(self): micro_cond_embed_dim=10, encoder_hidden_size=8, vocab_size=32, - codebook_size=32, # codebook size needs to be consistent with num_vq_embeddings for inpaint tests + codebook_size=32, in_channels=8, block_out_channels=8, num_res_blocks=1, @@ -69,19 +71,15 @@ def get_dummy_components(self): vqvae = VQModel( act_fn="silu", block_out_channels=[8], - down_block_types=[ - "DownEncoderBlock2D", - ], + down_block_types=["DownEncoderBlock2D"], in_channels=3, latent_channels=8, layers_per_block=1, norm_num_groups=8, - num_vq_embeddings=32, # reducing this to 16 or 8 -> RuntimeError: "cdist_cuda" not implemented for 'Half' + num_vq_embeddings=32, out_channels=3, sample_size=8, - up_block_types=[ - "UpDecoderBlock2D", - ], + up_block_types=["UpDecoderBlock2D"], mid_block_add_attention=False, lookup_from_codebook=True, ) @@ -100,7 +98,6 @@ def get_dummy_components(self): ) text_encoder = CLIPTextModelWithProjection(text_encoder_config) tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip") - components = { "transformer": transformer, "scheduler": scheduler, @@ -143,13 +140,11 @@ class AmusedInpaintPipelineSlowTests(unittest.TestCase): def test_amused_256(self): pipe = AmusedInpaintPipeline.from_pretrained("amused/amused-256") pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1.jpg") .resize((256, 256)) .convert("RGB") ) - mask_image = ( load_image( "https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1_mask.png" @@ -157,7 +152,6 @@ def test_amused_256(self): .resize((256, 256)) .convert("L") ) - image = pipe( "winter mountains", image, @@ -166,9 +160,7 @@ def test_amused_256(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) expected_slice = np.array([0.0699, 0.0716, 0.0608, 0.0715, 0.0797, 0.0638, 0.0802, 0.0924, 0.0634]) assert np.abs(image_slice - expected_slice).max() < 0.1 @@ -176,13 +168,11 @@ def test_amused_256(self): def test_amused_256_fp16(self): pipe = AmusedInpaintPipeline.from_pretrained("amused/amused-256", variant="fp16", torch_dtype=torch.float16) pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1.jpg") .resize((256, 256)) .convert("RGB") ) - mask_image = ( load_image( "https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1_mask.png" @@ -190,7 +180,6 @@ def test_amused_256_fp16(self): .resize((256, 256)) .convert("L") ) - image = pipe( "winter mountains", image, @@ -199,23 +188,19 @@ def test_amused_256_fp16(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 256, 256, 3) - expected_slice = np.array([0.0735, 0.0749, 0.0650, 0.0739, 0.0805, 0.0667, 0.0802, 0.0923, 0.0622]) + expected_slice = np.array([0.0735, 0.0749, 0.065, 0.0739, 0.0805, 0.0667, 0.0802, 0.0923, 0.0622]) assert np.abs(image_slice - expected_slice).max() < 0.1 def test_amused_512(self): pipe = AmusedInpaintPipeline.from_pretrained("amused/amused-512") pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1.jpg") .resize((512, 512)) .convert("RGB") ) - mask_image = ( load_image( "https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1_mask.png" @@ -223,7 +208,6 @@ def test_amused_512(self): .resize((512, 512)) .convert("L") ) - image = pipe( "winter mountains", image, @@ -232,9 +216,7 @@ def test_amused_512(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() - assert image.shape == (1, 512, 512, 3) expected_slice = np.array([0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0005, 0.0]) assert np.abs(image_slice - expected_slice).max() < 0.05 @@ -242,13 +224,11 @@ def test_amused_512(self): def test_amused_512_fp16(self): pipe = AmusedInpaintPipeline.from_pretrained("amused/amused-512", variant="fp16", torch_dtype=torch.float16) pipe.to(torch_device) - image = ( load_image("https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1.jpg") .resize((512, 512)) .convert("RGB") ) - mask_image = ( load_image( "https://huggingface.co/datasets/diffusers/docs-images/resolve/main/open_muse/mountains_1_mask.png" @@ -256,7 +236,6 @@ def test_amused_512_fp16(self): .resize((512, 512)) .convert("L") ) - image = pipe( "winter mountains", image, @@ -265,9 +244,8 @@ def test_amused_512_fp16(self): num_inference_steps=2, output_type="np", ).images - image_slice = image[0, -3:, -3:, -1].flatten() assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0025, 0.0]) - assert np.abs(image_slice - expected_slice).max() < 3e-3 + expected_slice = np.array([0.0227, 0.0157, 0.0098, 0.0213, 0.0250, 0.0127, 0.0280, 0.0380, 0.0095]) + assert np.abs(image_slice - expected_slice).max() < 0.003 diff --git a/tests/pipelines/controlnet/test_controlnet.py b/tests/pipelines/controlnet/test_controlnet.py index a2afc520944a..302712dbfd0d 100644 --- a/tests/pipelines/controlnet/test_controlnet.py +++ b/tests/pipelines/controlnet/test_controlnet.py @@ -73,7 +73,7 @@ def _test_stable_diffusion_compile(in_queue, out_queue, timeout): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.to("cuda") pipe.set_progress_bar_config(disable=None) @@ -715,7 +715,7 @@ def test_canny(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -742,7 +742,7 @@ def test_depth(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-depth") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -769,7 +769,7 @@ def test_hed(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-hed") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -796,7 +796,7 @@ def test_mlsd(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-mlsd") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -823,7 +823,7 @@ def test_normal(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-normal") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -850,7 +850,7 @@ def test_openpose(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-openpose") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -877,7 +877,7 @@ def test_scribble(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-scribble") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -904,7 +904,7 @@ def test_seg(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-seg") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -935,7 +935,7 @@ def test_sequential_cpu_offloading(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-seg") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.set_progress_bar_config(disable=None) pipe.enable_attention_slicing() @@ -961,7 +961,7 @@ def test_canny_guess_mode(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -993,7 +993,7 @@ def test_canny_guess_mode_euler(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config) pipe.enable_model_cpu_offload() @@ -1035,7 +1035,7 @@ def test_v11_shuffle_global_pool_conditions(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/control_v11e_sd15_shuffle") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -1081,7 +1081,7 @@ def test_pose_and_canny(self): controlnet_pose = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-openpose") pipe = StableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=[controlnet_pose, controlnet_canny] + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=[controlnet_pose, controlnet_canny] ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/controlnet/test_controlnet_img2img.py b/tests/pipelines/controlnet/test_controlnet_img2img.py index 05a484a3b8e1..2b22b3e5a76d 100644 --- a/tests/pipelines/controlnet/test_controlnet_img2img.py +++ b/tests/pipelines/controlnet/test_controlnet_img2img.py @@ -407,7 +407,7 @@ def test_canny(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetImg2ImgPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/controlnet/test_controlnet_inpaint.py b/tests/pipelines/controlnet/test_controlnet_inpaint.py index 10a88bf2845d..eba493c20588 100644 --- a/tests/pipelines/controlnet/test_controlnet_inpaint.py +++ b/tests/pipelines/controlnet/test_controlnet_inpaint.py @@ -459,7 +459,7 @@ def test_canny(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/sd-controlnet-canny") pipe = StableDiffusionControlNetInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None, controlnet=controlnet + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None, controlnet=controlnet ) pipe.enable_model_cpu_offload() pipe.set_progress_bar_config(disable=None) @@ -504,7 +504,7 @@ def test_inpaint(self): controlnet = ControlNetModel.from_pretrained("lllyasviel/control_v11p_sd15_inpaint") pipe = StableDiffusionControlNetInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, controlnet=controlnet + "Jiali/stable-diffusion-1.5", safety_checker=None, controlnet=controlnet ) pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config) pipe.enable_model_cpu_offload() diff --git a/tests/pipelines/controlnet/test_flax_controlnet.py b/tests/pipelines/controlnet/test_flax_controlnet.py index db19bd84a168..6ab66ccb6139 100644 --- a/tests/pipelines/controlnet/test_flax_controlnet.py +++ b/tests/pipelines/controlnet/test_flax_controlnet.py @@ -41,7 +41,7 @@ def test_canny(self): "lllyasviel/sd-controlnet-canny", from_pt=True, dtype=jnp.bfloat16 ) pipe, params = FlaxStableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", controlnet=controlnet, from_pt=True, dtype=jnp.bfloat16 + "Jiali/stable-diffusion-1.5", controlnet=controlnet, from_pt=True, dtype=jnp.bfloat16 ) params["controlnet"] = controlnet_params @@ -86,7 +86,7 @@ def test_pose(self): "lllyasviel/sd-controlnet-openpose", from_pt=True, dtype=jnp.bfloat16 ) pipe, params = FlaxStableDiffusionControlNetPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", controlnet=controlnet, from_pt=True, dtype=jnp.bfloat16 + "Jiali/stable-diffusion-1.5", controlnet=controlnet, from_pt=True, dtype=jnp.bfloat16 ) params["controlnet"] = controlnet_params diff --git a/tests/pipelines/ip_adapters/test_ip_adapter_stable_diffusion.py b/tests/pipelines/ip_adapters/test_ip_adapter_stable_diffusion.py index 694a4d4574bf..741639e2d09e 100644 --- a/tests/pipelines/ip_adapters/test_ip_adapter_stable_diffusion.py +++ b/tests/pipelines/ip_adapters/test_ip_adapter_stable_diffusion.py @@ -170,7 +170,7 @@ class IPAdapterSDIntegrationTests(IPAdapterNightlyTestsMixin): def test_text_to_image(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin") @@ -200,7 +200,7 @@ def test_text_to_image(self): def test_image_to_image(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionImg2ImgPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin") @@ -232,7 +232,7 @@ def test_image_to_image(self): def test_inpainting(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin") @@ -260,7 +260,7 @@ def test_inpainting(self): def test_text_to_image_model_cpu_offload(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter_sd15.bin") pipeline.to(torch_device) @@ -287,7 +287,7 @@ def test_text_to_image_model_cpu_offload(self): def test_text_to_image_full_face(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter("h94/IP-Adapter", subfolder="models", weight_name="ip-adapter-full-face_sd15.bin") @@ -304,7 +304,7 @@ def test_text_to_image_full_face(self): def test_unload(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) before_processors = [attn_proc.__class__ for attn_proc in pipeline.unet.attn_processors.values()] pipeline.to(torch_device) @@ -323,7 +323,7 @@ def test_unload(self): def test_multi(self): image_encoder = self.get_image_encoder(repo_id="h94/IP-Adapter", subfolder="models/image_encoder") pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", image_encoder=image_encoder, safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter( @@ -343,7 +343,7 @@ def test_multi(self): def test_text_to_image_face_id(self): pipeline = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, torch_dtype=self.dtype + "Jiali/stable-diffusion-1.5", safety_checker=None, torch_dtype=self.dtype ) pipeline.to(torch_device) pipeline.load_ip_adapter( diff --git a/tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion.py b/tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion.py index 26417768843c..12742def67f8 100644 --- a/tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion.py +++ b/tests/pipelines/ledits_pp/test_ledits_pp_stable_diffusion.py @@ -224,7 +224,7 @@ def setUpClass(cls): def test_ledits_pp_editing(self): pipe = LEditsPPPipelineStableDiffusion.from_pretrained( - "runwayml/stable-diffusion-v1-5", safety_checker=None, torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", safety_checker=None, torch_dtype=torch.float16 ) pipe = pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/pag/test_pag_pixart_sigma.py b/tests/pipelines/pag/test_pag_pixart_sigma.py index 70b528dede56..7de19e0f00fc 100644 --- a/tests/pipelines/pag/test_pag_pixart_sigma.py +++ b/tests/pipelines/pag/test_pag_pixart_sigma.py @@ -33,7 +33,6 @@ from diffusers.utils.testing_utils import ( CaptureLogger, enable_full_determinism, - print_tensor_test, torch_device, ) @@ -173,7 +172,6 @@ def test_pag_inference(self): inputs = self.get_dummy_inputs(device) image = pipe_pag(**inputs).images image_slice = image[0, -3:, -3:, -1] - print_tensor_test(image_slice) assert image.shape == ( 1, diff --git a/tests/pipelines/pag/test_pag_sd.py b/tests/pipelines/pag/test_pag_sd.py index e9adb3ac447e..6a644e02f5e8 100644 --- a/tests/pipelines/pag/test_pag_sd.py +++ b/tests/pipelines/pag/test_pag_sd.py @@ -283,7 +283,7 @@ def test_pag_inference(self): @require_torch_gpu class StableDiffusionPAGPipelineIntegrationTests(unittest.TestCase): pipeline_class = StableDiffusionPAGPipeline - repo_id = "runwayml/stable-diffusion-v1-5" + repo_id = "Jiali/stable-diffusion-1.5" def setUp(self): super().setUp() diff --git a/tests/pipelines/semantic_stable_diffusion/test_semantic_diffusion.py b/tests/pipelines/semantic_stable_diffusion/test_semantic_diffusion.py index 1cc3111c2631..3a9d3815e72d 100644 --- a/tests/pipelines/semantic_stable_diffusion/test_semantic_diffusion.py +++ b/tests/pipelines/semantic_stable_diffusion/test_semantic_diffusion.py @@ -287,7 +287,7 @@ def tearDown(self): def test_positive_guidance(self): torch_device = "cuda" - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5") pipe = pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -370,7 +370,7 @@ def test_positive_guidance(self): def test_negative_guidance(self): torch_device = "cuda" - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5") pipe = pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -453,7 +453,7 @@ def test_negative_guidance(self): def test_multi_cond_guidance(self): torch_device = "cuda" - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5") pipe = pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -536,7 +536,7 @@ def test_multi_cond_guidance(self): def test_guidance_fp16(self): torch_device = "cuda" - pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16) + pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16) pipe = pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion.py b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion.py index c3bfa6b5dabc..24ea6e07280e 100644 --- a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion.py +++ b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion.py @@ -250,10 +250,10 @@ def test_inference_default_pndm(self): def test_inference_ddim(self): ddim_scheduler = DDIMScheduler.from_pretrained( - "runwayml/stable-diffusion-v1-5", subfolder="scheduler", revision="onnx" + "Jiali/stable-diffusion-1.5", subfolder="scheduler", revision="onnx" ) sd_pipe = OnnxStableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", revision="onnx", scheduler=ddim_scheduler, safety_checker=None, @@ -276,10 +276,10 @@ def test_inference_ddim(self): def test_inference_k_lms(self): lms_scheduler = LMSDiscreteScheduler.from_pretrained( - "runwayml/stable-diffusion-v1-5", subfolder="scheduler", revision="onnx" + "Jiali/stable-diffusion-1.5", subfolder="scheduler", revision="onnx" ) sd_pipe = OnnxStableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", revision="onnx", scheduler=lms_scheduler, safety_checker=None, @@ -327,7 +327,7 @@ def test_callback_fn(step: int, timestep: int, latents: np.ndarray) -> None: test_callback_fn.has_been_called = False pipe = OnnxStableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", revision="onnx", safety_checker=None, feature_extractor=None, @@ -352,7 +352,7 @@ def test_callback_fn(step: int, timestep: int, latents: np.ndarray) -> None: def test_stable_diffusion_no_safety_checker(self): pipe = OnnxStableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", revision="onnx", safety_checker=None, feature_extractor=None, diff --git a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_img2img.py b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_img2img.py index 274cb6718233..086354a2a649 100644 --- a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_img2img.py +++ b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_img2img.py @@ -210,10 +210,10 @@ def test_inference_k_lms(self): ) init_image = init_image.resize((768, 512)) lms_scheduler = LMSDiscreteScheduler.from_pretrained( - "runwayml/stable-diffusion-v1-5", subfolder="scheduler", revision="onnx" + "Jiali/stable-diffusion-1.5", subfolder="scheduler", revision="onnx" ) pipe = OnnxStableDiffusionImg2ImgPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", revision="onnx", scheduler=lms_scheduler, safety_checker=None, diff --git a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_inpaint.py b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_inpaint.py index 6426547bae5b..09048b5c0e0f 100644 --- a/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_inpaint.py +++ b/tests/pipelines/stable_diffusion/test_onnx_stable_diffusion_inpaint.py @@ -68,7 +68,7 @@ def test_inference_default_pndm(self): "/in_paint/overture-creations-5sI6fQgYIuo_mask.png" ) pipe = OnnxStableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", + "botp/stable-diffusion-v1-5-inpainting", revision="onnx", safety_checker=None, feature_extractor=None, @@ -107,10 +107,10 @@ def test_inference_k_lms(self): "/in_paint/overture-creations-5sI6fQgYIuo_mask.png" ) lms_scheduler = LMSDiscreteScheduler.from_pretrained( - "runwayml/stable-diffusion-inpainting", subfolder="scheduler", revision="onnx" + "botp/stable-diffusion-v1-5-inpainting", subfolder="scheduler", revision="onnx" ) pipe = OnnxStableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", + "botp/stable-diffusion-v1-5-inpainting", revision="onnx", scheduler=lms_scheduler, safety_checker=None, diff --git a/tests/pipelines/stable_diffusion/test_stable_diffusion.py b/tests/pipelines/stable_diffusion/test_stable_diffusion.py index fea63f2b11da..5bb13fac9b78 100644 --- a/tests/pipelines/stable_diffusion/test_stable_diffusion.py +++ b/tests/pipelines/stable_diffusion/test_stable_diffusion.py @@ -1332,7 +1332,7 @@ def tearDown(self): def test_download_from_hub(self): ckpt_paths = [ - "https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/v1-5-pruned-emaonly.safetensors", + "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors", "https://huggingface.co/WarriorMama777/OrangeMixs/blob/main/Models/AbyssOrangeMix/AbyssOrangeMix.safetensors", ] @@ -1346,8 +1346,8 @@ def test_download_from_hub(self): assert image_out.shape == (512, 512, 3) def test_download_local(self): - ckpt_filename = hf_hub_download("runwayml/stable-diffusion-v1-5", filename="v1-5-pruned-emaonly.safetensors") - config_filename = hf_hub_download("runwayml/stable-diffusion-v1-5", filename="v1-inference.yaml") + ckpt_filename = hf_hub_download("Jiali/stable-diffusion-1.5", filename="v1-5-pruned-emaonly.safetensors") + config_filename = hf_hub_download("Jiali/stable-diffusion-1.5", filename="v1-inference.yaml") pipe = StableDiffusionPipeline.from_single_file( ckpt_filename, config_files={"v1": config_filename}, torch_dtype=torch.float16 @@ -1402,7 +1402,7 @@ def test_stable_diffusion_1_4_pndm(self): assert max_diff < 1e-3 def test_stable_diffusion_1_5_pndm(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5").to(torch_device) + sd_pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5").to(torch_device) sd_pipe.set_progress_bar_config(disable=None) inputs = self.get_inputs(torch_device) @@ -1483,9 +1483,9 @@ def get_inputs(self, generator_device="cpu", seed=0): return inputs def get_pipeline_output_without_device_map(self): - sd_pipe = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", torch_dtype=torch.float16 - ).to(torch_device) + sd_pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", torch_dtype=torch.float16).to( + torch_device + ) sd_pipe.set_progress_bar_config(disable=True) inputs = self.get_inputs() no_device_map_image = sd_pipe(**inputs).images @@ -1498,7 +1498,7 @@ def test_forward_pass_balanced_device_map(self): no_device_map_image = self.get_pipeline_output_without_device_map() sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) sd_pipe_with_device_map.set_progress_bar_config(disable=True) inputs = self.get_inputs() @@ -1509,7 +1509,7 @@ def test_forward_pass_balanced_device_map(self): def test_components_put_in_right_devices(self): sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) assert len(set(sd_pipe_with_device_map.hf_device_map.values())) >= 2 @@ -1518,7 +1518,7 @@ def test_max_memory(self): no_device_map_image = self.get_pipeline_output_without_device_map() sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", + "Jiali/stable-diffusion-1.5", device_map="balanced", max_memory={0: "1GB", 1: "1GB"}, torch_dtype=torch.float16, @@ -1532,7 +1532,7 @@ def test_max_memory(self): def test_reset_device_map(self): sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) sd_pipe_with_device_map.reset_device_map() @@ -1544,7 +1544,7 @@ def test_reset_device_map(self): def test_reset_device_map_to(self): sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) sd_pipe_with_device_map.reset_device_map() @@ -1556,7 +1556,7 @@ def test_reset_device_map_to(self): def test_reset_device_map_enable_model_cpu_offload(self): sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) sd_pipe_with_device_map.reset_device_map() @@ -1568,7 +1568,7 @@ def test_reset_device_map_enable_model_cpu_offload(self): def test_reset_device_map_enable_sequential_cpu_offload(self): sd_pipe_with_device_map = StableDiffusionPipeline.from_pretrained( - "runwayml/stable-diffusion-v1-5", device_map="balanced", torch_dtype=torch.float16 + "Jiali/stable-diffusion-1.5", device_map="balanced", torch_dtype=torch.float16 ) sd_pipe_with_device_map.reset_device_map() diff --git a/tests/pipelines/stable_diffusion/test_stable_diffusion_img2img.py b/tests/pipelines/stable_diffusion/test_stable_diffusion_img2img.py index 64ebf641d1e6..e6de84781f26 100644 --- a/tests/pipelines/stable_diffusion/test_stable_diffusion_img2img.py +++ b/tests/pipelines/stable_diffusion/test_stable_diffusion_img2img.py @@ -566,7 +566,7 @@ def test_stable_diffusion_pipeline_with_model_offloading(self): assert module.device == torch.device("cpu") def test_img2img_2nd_order(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.scheduler = HeunDiscreteScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -630,7 +630,7 @@ def test_stable_diffusion_img2img_pipeline_multiple_of_8(self): assert np.abs(image_slice.flatten() - expected_slice).max() < 5e-3 def test_img2img_safety_checker_works(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -686,7 +686,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 return inputs def test_img2img_pndm(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -701,7 +701,7 @@ def test_img2img_pndm(self): assert max_diff < 1e-3 def test_img2img_ddim(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.scheduler = DDIMScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -717,7 +717,7 @@ def test_img2img_ddim(self): assert max_diff < 1e-3 def test_img2img_lms(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.scheduler = LMSDiscreteScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -733,7 +733,7 @@ def test_img2img_lms(self): assert max_diff < 1e-3 def test_img2img_dpm(self): - sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe.scheduler = DPMSolverMultistepScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/stable_diffusion/test_stable_diffusion_inpaint.py b/tests/pipelines/stable_diffusion/test_stable_diffusion_inpaint.py index 6965954f7eb3..28da97be9362 100644 --- a/tests/pipelines/stable_diffusion/test_stable_diffusion_inpaint.py +++ b/tests/pipelines/stable_diffusion/test_stable_diffusion_inpaint.py @@ -76,7 +76,7 @@ def _test_inpaint_compile(in_queue, out_queue, timeout): inputs["generator"] = torch.Generator(device=torch_device).manual_seed(seed) pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.scheduler = PNDMScheduler.from_config(pipe.scheduler.config) @@ -628,7 +628,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 def test_stable_diffusion_inpaint_ddim(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -645,7 +645,7 @@ def test_stable_diffusion_inpaint_ddim(self): def test_stable_diffusion_inpaint_fp16(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", torch_dtype=torch.float16, safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", torch_dtype=torch.float16, safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.to(torch_device) @@ -662,7 +662,7 @@ def test_stable_diffusion_inpaint_fp16(self): def test_stable_diffusion_inpaint_pndm(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.scheduler = PNDMScheduler.from_config(pipe.scheduler.config) pipe.to(torch_device) @@ -680,7 +680,7 @@ def test_stable_diffusion_inpaint_pndm(self): def test_stable_diffusion_inpaint_k_lms(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) pipe.to(torch_device) @@ -702,7 +702,7 @@ def test_stable_diffusion_inpaint_with_sequential_cpu_offloading(self): torch.cuda.reset_peak_memory_stats() pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None, torch_dtype=torch.float16 + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None, torch_dtype=torch.float16 ) pipe.set_progress_bar_config(disable=None) pipe.enable_attention_slicing(1) @@ -728,7 +728,7 @@ def test_inpaint_compile(self): def test_stable_diffusion_inpaint_pil_input_resolution_test(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) pipe.to(torch_device) @@ -747,7 +747,7 @@ def test_stable_diffusion_inpaint_pil_input_resolution_test(self): def test_stable_diffusion_inpaint_strength_test(self): pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) pipe.unet.set_default_attn_processor() @@ -767,7 +767,7 @@ def test_stable_diffusion_inpaint_strength_test(self): assert np.abs(expected_slice - image_slice).max() < 1e-3 def test_stable_diffusion_simple_inpaint_ddim(self): - pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None) + pipe = StableDiffusionInpaintPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None) pipe.unet.set_default_attn_processor() pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -818,7 +818,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 def test_stable_diffusion_inpaint_ddim(self): vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-1-5") pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.vae = vae pipe.unet.set_default_attn_processor() @@ -840,7 +840,7 @@ def test_stable_diffusion_inpaint_fp16(self): "cross-attention/asymmetric-autoencoder-kl-x-1-5", torch_dtype=torch.float16 ) pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", torch_dtype=torch.float16, safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", torch_dtype=torch.float16, safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.vae = vae @@ -860,7 +860,7 @@ def test_stable_diffusion_inpaint_fp16(self): def test_stable_diffusion_inpaint_pndm(self): vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-1-5") pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.vae = vae @@ -881,7 +881,7 @@ def test_stable_diffusion_inpaint_pndm(self): def test_stable_diffusion_inpaint_k_lms(self): vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-1-5") pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.vae = vae @@ -906,7 +906,7 @@ def test_stable_diffusion_inpaint_with_sequential_cpu_offloading(self): "cross-attention/asymmetric-autoencoder-kl-x-1-5", torch_dtype=torch.float16 ) pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None, torch_dtype=torch.float16 + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None, torch_dtype=torch.float16 ) pipe.vae = vae pipe.set_progress_bar_config(disable=None) @@ -930,7 +930,7 @@ def test_stable_diffusion_inpaint_pil_input_resolution_test(self): "cross-attention/asymmetric-autoencoder-kl-x-1-5", ) pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.vae = vae pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) @@ -951,7 +951,7 @@ def test_stable_diffusion_inpaint_pil_input_resolution_test(self): def test_stable_diffusion_inpaint_strength_test(self): vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-1-5") pipe = StableDiffusionInpaintPipeline.from_pretrained( - "runwayml/stable-diffusion-inpainting", safety_checker=None + "botp/stable-diffusion-v1-5-inpainting", safety_checker=None ) pipe.unet.set_default_attn_processor() pipe.vae = vae @@ -973,7 +973,7 @@ def test_stable_diffusion_inpaint_strength_test(self): def test_stable_diffusion_simple_inpaint_ddim(self): vae = AsymmetricAutoencoderKL.from_pretrained("cross-attention/asymmetric-autoencoder-kl-x-1-5") - pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None) + pipe = StableDiffusionInpaintPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None) pipe.vae = vae pipe.unet.set_default_attn_processor() pipe.to(torch_device) @@ -993,7 +993,7 @@ def test_download_local(self): vae = AsymmetricAutoencoderKL.from_pretrained( "cross-attention/asymmetric-autoencoder-kl-x-1-5", torch_dtype=torch.float16 ) - filename = hf_hub_download("runwayml/stable-diffusion-inpainting", filename="sd-v1-5-inpainting.ckpt") + filename = hf_hub_download("botp/stable-diffusion-v1-5-inpainting", filename="sd-v1-5-inpainting.ckpt") pipe = StableDiffusionInpaintPipeline.from_single_file(filename, torch_dtype=torch.float16) pipe.vae = vae @@ -1042,7 +1042,7 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 return inputs def test_inpaint_ddim(self): - sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-inpainting") + sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("botp/stable-diffusion-v1-5-inpainting") sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -1057,7 +1057,7 @@ def test_inpaint_ddim(self): assert max_diff < 1e-3 def test_inpaint_pndm(self): - sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-inpainting") + sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("botp/stable-diffusion-v1-5-inpainting") sd_pipe.scheduler = PNDMScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -1073,7 +1073,7 @@ def test_inpaint_pndm(self): assert max_diff < 1e-3 def test_inpaint_lms(self): - sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-inpainting") + sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("botp/stable-diffusion-v1-5-inpainting") sd_pipe.scheduler = LMSDiscreteScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -1089,7 +1089,7 @@ def test_inpaint_lms(self): assert max_diff < 1e-3 def test_inpaint_dpm(self): - sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("runwayml/stable-diffusion-inpainting") + sd_pipe = StableDiffusionInpaintPipeline.from_pretrained("botp/stable-diffusion-v1-5-inpainting") sd_pipe.scheduler = DPMSolverMultistepScheduler.from_config(sd_pipe.scheduler.config) sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/stable_diffusion_2/test_stable_diffusion.py b/tests/pipelines/stable_diffusion_2/test_stable_diffusion.py index c39bfb1984b5..e7114d19e208 100644 --- a/tests/pipelines/stable_diffusion_2/test_stable_diffusion.py +++ b/tests/pipelines/stable_diffusion_2/test_stable_diffusion.py @@ -23,7 +23,6 @@ from diffusers import ( AutoencoderKL, DDIMScheduler, - DPMSolverMultistepScheduler, EulerAncestralDiscreteScheduler, EulerDiscreteScheduler, LMSDiscreteScheduler, @@ -353,34 +352,6 @@ def test_stable_diffusion_default_ddim(self): expected_slice = np.array([0.49493, 0.47896, 0.40798, 0.54214, 0.53212, 0.48202, 0.47656, 0.46329, 0.48506]) assert np.abs(image_slice - expected_slice).max() < 7e-3 - def test_stable_diffusion_pndm(self): - pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-base") - pipe.scheduler = PNDMScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = pipe(**inputs).images - image_slice = image[0, -3:, -3:, -1].flatten() - - assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.49493, 0.47896, 0.40798, 0.54214, 0.53212, 0.48202, 0.47656, 0.46329, 0.48506]) - assert np.abs(image_slice - expected_slice).max() < 7e-3 - - def test_stable_diffusion_k_lms(self): - pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-base") - pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = pipe(**inputs).images - image_slice = image[0, -3:, -3:, -1].flatten() - - assert image.shape == (1, 512, 512, 3) - expected_slice = np.array([0.10440, 0.13115, 0.11100, 0.10141, 0.11440, 0.07215, 0.11332, 0.09693, 0.10006]) - assert np.abs(image_slice - expected_slice).max() < 3e-3 - @require_torch_gpu def test_stable_diffusion_attention_slicing(self): torch.cuda.reset_peak_memory_stats() @@ -413,124 +384,6 @@ def test_stable_diffusion_attention_slicing(self): max_diff = numpy_cosine_similarity_distance(image.flatten(), image_sliced.flatten()) assert max_diff < 5e-3 - def test_stable_diffusion_text2img_intermediate_state(self): - number_of_steps = 0 - - def callback_fn(step: int, timestep: int, latents: torch.Tensor) -> None: - callback_fn.has_been_called = True - nonlocal number_of_steps - number_of_steps += 1 - if step == 1: - latents = latents.detach().cpu().numpy() - assert latents.shape == (1, 4, 64, 64) - latents_slice = latents[0, -3:, -3:, -1] - expected_slice = np.array( - [-0.3862, -0.4507, -1.1729, 0.0686, -1.1045, 0.7124, -1.8301, 0.1903, 1.2773] - ) - - assert np.abs(latents_slice.flatten() - expected_slice).max() < 5e-2 - elif step == 2: - latents = latents.detach().cpu().numpy() - assert latents.shape == (1, 4, 64, 64) - latents_slice = latents[0, -3:, -3:, -1] - expected_slice = np.array( - [0.2720, -0.1863, -0.7383, -0.5029, -0.7534, 0.3970, -0.7646, 0.4468, 1.2686] - ) - - assert np.abs(latents_slice.flatten() - expected_slice).max() < 5e-2 - - callback_fn.has_been_called = False - - pipe = StableDiffusionPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-base", torch_dtype=torch.float16 - ) - pipe = pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - inputs = self.get_inputs(torch_device, dtype=torch.float16) - pipe(**inputs, callback=callback_fn, callback_steps=1) - assert callback_fn.has_been_called - assert number_of_steps == inputs["num_inference_steps"] - - @require_torch_gpu - def test_stable_diffusion_pipeline_with_sequential_cpu_offloading(self): - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - pipe = StableDiffusionPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-base", torch_dtype=torch.float16 - ) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing(1) - pipe.enable_sequential_cpu_offload() - - inputs = self.get_inputs(torch_device, dtype=torch.float16) - _ = pipe(**inputs) - - mem_bytes = torch.cuda.max_memory_allocated() - # make sure that less than 2.8 GB is allocated - assert mem_bytes < 2.8 * 10**9 - - @require_torch_gpu - def test_stable_diffusion_pipeline_with_model_offloading(self): - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - inputs = self.get_inputs(torch_device, dtype=torch.float16) - - # Normal inference - - pipe = StableDiffusionPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-base", - torch_dtype=torch.float16, - ) - pipe.unet.set_default_attn_processor() - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - outputs = pipe(**inputs) - mem_bytes = torch.cuda.max_memory_allocated() - - # With model offloading - - # Reload but don't move to cuda - pipe = StableDiffusionPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-base", - torch_dtype=torch.float16, - ) - pipe.unet.set_default_attn_processor() - - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - pipe.enable_model_cpu_offload() - pipe.set_progress_bar_config(disable=None) - inputs = self.get_inputs(torch_device, dtype=torch.float16) - outputs_offloaded = pipe(**inputs) - mem_bytes_offloaded = torch.cuda.max_memory_allocated() - - images = outputs.images - images_offloaded = outputs_offloaded.images - max_diff = numpy_cosine_similarity_distance(images.flatten(), images_offloaded.flatten()) - assert max_diff < 1e-3 - assert mem_bytes_offloaded < mem_bytes - assert mem_bytes_offloaded < 3 * 10**9 - for module in pipe.text_encoder, pipe.unet, pipe.vae: - assert module.device == torch.device("cpu") - - # With attention slicing - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - pipe.enable_attention_slicing() - _ = pipe(**inputs) - mem_bytes_slicing = torch.cuda.max_memory_allocated() - assert mem_bytes_slicing < mem_bytes_offloaded - @nightly @require_torch_accelerator @@ -554,99 +407,22 @@ def get_inputs(self, device, generator_device="cpu", dtype=torch.float32, seed=0 "prompt": "a photograph of an astronaut riding a horse", "latents": latents, "generator": generator, - "num_inference_steps": 50, + "num_inference_steps": 2, "guidance_scale": 7.5, "output_type": "np", } return inputs - def test_stable_diffusion_2_0_default_ddim(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-base").to(torch_device) - sd_pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = sd_pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_0_base_ddim.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_stable_diffusion_2_1_default_pndm(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base").to(torch_device) - sd_pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = sd_pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_1_base_pndm.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_stable_diffusion_ddim(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base").to(torch_device) - sd_pipe.scheduler = DDIMScheduler.from_config(sd_pipe.scheduler.config) - sd_pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = sd_pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_1_base_ddim.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_stable_diffusion_lms(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base").to(torch_device) - sd_pipe.scheduler = LMSDiscreteScheduler.from_config(sd_pipe.scheduler.config) - sd_pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = sd_pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_1_base_lms.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_stable_diffusion_euler(self): + def test_stable_diffusion_2_1_default(self): sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base").to(torch_device) - sd_pipe.scheduler = EulerDiscreteScheduler.from_config(sd_pipe.scheduler.config) - sd_pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs(torch_device) - image = sd_pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_1_base_euler.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_stable_diffusion_dpm(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-2-1-base").to(torch_device) - sd_pipe.scheduler = DPMSolverMultistepScheduler.from_config( - sd_pipe.scheduler.config, final_sigmas_type="sigma_min" - ) sd_pipe.set_progress_bar_config(disable=None) inputs = self.get_inputs(torch_device) - inputs["num_inference_steps"] = 25 image = sd_pipe(**inputs).images[0] expected_image = load_numpy( "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_2_text2img/stable_diffusion_2_1_base_dpm_multi.npy" + "/stable_diffusion_2_text2img/stable_diffusion_2_0_pndm.npy" ) max_diff = np.abs(expected_image - image).max() assert max_diff < 1e-3 diff --git a/tests/pipelines/stable_diffusion_2/test_stable_diffusion_depth.py b/tests/pipelines/stable_diffusion_2/test_stable_diffusion_depth.py index 9a3a93acd6ce..42eef061069e 100644 --- a/tests/pipelines/stable_diffusion_2/test_stable_diffusion_depth.py +++ b/tests/pipelines/stable_diffusion_2/test_stable_diffusion_depth.py @@ -32,9 +32,6 @@ from diffusers import ( AutoencoderKL, - DDIMScheduler, - DPMSolverMultistepScheduler, - LMSDiscreteScheduler, PNDMScheduler, StableDiffusionDepth2ImgPipeline, UNet2DConditionModel, @@ -416,102 +413,6 @@ def test_stable_diffusion_depth2img_pipeline_default(self): assert np.abs(expected_slice - image_slice).max() < 6e-1 - def test_stable_diffusion_depth2img_pipeline_k_lms(self): - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-depth", safety_checker=None - ) - pipe.unet.set_default_attn_processor() - pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - inputs = self.get_inputs() - image = pipe(**inputs).images - image_slice = image[0, 253:256, 253:256, -1].flatten() - - assert image.shape == (1, 480, 640, 3) - expected_slice = np.array([0.6363, 0.6274, 0.6309, 0.6370, 0.6226, 0.6286, 0.6213, 0.6453, 0.6306]) - - assert np.abs(expected_slice - image_slice).max() < 8e-4 - - def test_stable_diffusion_depth2img_pipeline_ddim(self): - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-depth", safety_checker=None - ) - pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - inputs = self.get_inputs() - image = pipe(**inputs).images - image_slice = image[0, 253:256, 253:256, -1].flatten() - - assert image.shape == (1, 480, 640, 3) - expected_slice = np.array([0.6424, 0.6524, 0.6249, 0.6041, 0.6634, 0.6420, 0.6522, 0.6555, 0.6436]) - - assert np.abs(expected_slice - image_slice).max() < 5e-4 - - def test_stable_diffusion_depth2img_intermediate_state(self): - number_of_steps = 0 - - def callback_fn(step: int, timestep: int, latents: torch.Tensor) -> None: - callback_fn.has_been_called = True - nonlocal number_of_steps - number_of_steps += 1 - if step == 1: - latents = latents.detach().cpu().numpy() - assert latents.shape == (1, 4, 60, 80) - latents_slice = latents[0, -3:, -3:, -1] - expected_slice = np.array( - [-0.7168, -1.5137, -0.1418, -2.9219, -2.7266, -2.4414, -2.1035, -3.0078, -1.7051] - ) - - assert np.abs(latents_slice.flatten() - expected_slice).max() < 5e-2 - elif step == 2: - latents = latents.detach().cpu().numpy() - assert latents.shape == (1, 4, 60, 80) - latents_slice = latents[0, -3:, -3:, -1] - expected_slice = np.array( - [-0.7109, -1.5068, -0.1403, -2.9160, -2.7207, -2.4414, -2.1035, -3.0059, -1.7090] - ) - - assert np.abs(latents_slice.flatten() - expected_slice).max() < 5e-2 - - callback_fn.has_been_called = False - - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-depth", safety_checker=None, torch_dtype=torch.float16 - ) - pipe = pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - inputs = self.get_inputs(dtype=torch.float16) - pipe(**inputs, callback=callback_fn, callback_steps=1) - assert callback_fn.has_been_called - assert number_of_steps == 2 - - def test_stable_diffusion_pipeline_with_sequential_cpu_offloading(self): - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained( - "stabilityai/stable-diffusion-2-depth", safety_checker=None, torch_dtype=torch.float16 - ) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing(1) - pipe.enable_sequential_cpu_offload() - - inputs = self.get_inputs(dtype=torch.float16) - _ = pipe(**inputs) - - mem_bytes = torch.cuda.max_memory_allocated() - # make sure that less than 2.9 GB is allocated - assert mem_bytes < 2.9 * 10**9 - @nightly @require_torch_gpu @@ -535,14 +436,14 @@ def get_inputs(self, device="cpu", dtype=torch.float32, seed=0): "prompt": "two tigers", "image": init_image, "generator": generator, - "num_inference_steps": 3, + "num_inference_steps": 2, "strength": 0.75, "guidance_scale": 7.5, "output_type": "np", } return inputs - def test_depth2img_pndm(self): + def test_depth2img(self): pipe = StableDiffusionDepth2ImgPipeline.from_pretrained("stabilityai/stable-diffusion-2-depth") pipe.to(torch_device) pipe.set_progress_bar_config(disable=None) @@ -556,52 +457,3 @@ def test_depth2img_pndm(self): ) max_diff = np.abs(expected_image - image).max() assert max_diff < 1e-3 - - def test_depth2img_ddim(self): - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained("stabilityai/stable-diffusion-2-depth") - pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs() - image = pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_depth2img/stable_diffusion_2_0_ddim.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_img2img_lms(self): - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained("stabilityai/stable-diffusion-2-depth") - pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs() - image = pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_depth2img/stable_diffusion_2_0_lms.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 - - def test_img2img_dpm(self): - pipe = StableDiffusionDepth2ImgPipeline.from_pretrained("stabilityai/stable-diffusion-2-depth") - pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - - inputs = self.get_inputs() - inputs["num_inference_steps"] = 30 - image = pipe(**inputs).images[0] - - expected_image = load_numpy( - "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main" - "/stable_diffusion_depth2img/stable_diffusion_2_0_dpm_multi.npy" - ) - max_diff = np.abs(expected_image - image).max() - assert max_diff < 1e-3 diff --git a/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3.py b/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3.py index 75a7d88ea4f2..94a85a56f510 100644 --- a/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3.py +++ b/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3.py @@ -265,18 +265,37 @@ def test_sd3_inference(self): image_slice = image[0, :10, :10] expected_slice = np.array( [ - [0.36132812, 0.30004883, 0.25830078], - [0.36669922, 0.31103516, 0.23754883], - [0.34814453, 0.29248047, 0.23583984], - [0.35791016, 0.30981445, 0.23999023], - [0.36328125, 0.31274414, 0.2607422], - [0.37304688, 0.32177734, 0.26171875], - [0.3671875, 0.31933594, 0.25756836], - [0.36035156, 0.31103516, 0.2578125], - [0.3857422, 0.33789062, 0.27563477], - [0.3701172, 0.31982422, 0.265625], - ], - dtype=np.float32, + 0.4648, + 0.4404, + 0.4177, + 0.5063, + 0.4800, + 0.4287, + 0.5425, + 0.5190, + 0.4717, + 0.5430, + 0.5195, + 0.4766, + 0.5361, + 0.5122, + 0.4612, + 0.4871, + 0.4749, + 0.4058, + 0.4756, + 0.4678, + 0.3804, + 0.4832, + 0.4822, + 0.3799, + 0.5103, + 0.5034, + 0.3953, + 0.5073, + 0.4839, + 0.3884, + ] ) max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), image_slice.flatten()) diff --git a/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_img2img.py b/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_img2img.py index dc68cc3ecdbd..9d131b28c308 100644 --- a/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_img2img.py +++ b/tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_img2img.py @@ -238,20 +238,39 @@ def test_sd3_img2img_inference(self): image_slice = image[0, :10, :10] expected_slice = np.array( [ - [0.50097656, 0.44726562, 0.40429688], - [0.5048828, 0.45703125, 0.38110352], - [0.4987793, 0.45141602, 0.38134766], - [0.49682617, 0.45336914, 0.38354492], - [0.49804688, 0.4555664, 0.39379883], - [0.5083008, 0.4645996, 0.40039062], - [0.50341797, 0.46240234, 0.39770508], - [0.49926758, 0.4572754, 0.39575195], - [0.50634766, 0.46435547, 0.39794922], - [0.50341797, 0.4572754, 0.39746094], - ], - dtype=np.float32, + 0.5435, + 0.4673, + 0.5732, + 0.4438, + 0.3557, + 0.4912, + 0.4331, + 0.3491, + 0.4915, + 0.4287, + 0.3477, + 0.4849, + 0.4355, + 0.3469, + 0.4871, + 0.4431, + 0.3538, + 0.4912, + 0.4521, + 0.3643, + 0.5059, + 0.4587, + 0.3730, + 0.5166, + 0.4685, + 0.3845, + 0.5264, + 0.4746, + 0.3914, + 0.5342, + ] ) max_diff = numpy_cosine_similarity_distance(expected_slice.flatten(), image_slice.flatten()) - assert max_diff < 1e-4, f"Outputs are not close enough, got {image_slice}" + assert max_diff < 1e-4, f"Outputs are not close enough, got {max_diff}" diff --git a/tests/pipelines/stable_diffusion_adapter/test_stable_diffusion_adapter.py b/tests/pipelines/stable_diffusion_adapter/test_stable_diffusion_adapter.py index 678bffcefaf0..8e0841f064e0 100644 --- a/tests/pipelines/stable_diffusion_adapter/test_stable_diffusion_adapter.py +++ b/tests/pipelines/stable_diffusion_adapter/test_stable_diffusion_adapter.py @@ -607,175 +607,14 @@ def tearDown(self): gc.collect() torch.cuda.empty_cache() - def test_stable_diffusion_adapter_color(self): - adapter_model = "TencentARC/t2iadapter_color_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "snail" - image_url = ( - "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/color.png" - ) - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_color_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_depth(self): - adapter_model = "TencentARC/t2iadapter_depth_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "snail" - image_url = ( - "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/color.png" - ) - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_color_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_depth_sd_v14(self): - adapter_model = "TencentARC/t2iadapter_depth_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "desk" - image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/desk_depth.png" - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_depth_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - def test_stable_diffusion_adapter_depth_sd_v15(self): adapter_model = "TencentARC/t2iadapter_depth_sd15v2" - sd_model = "runwayml/stable-diffusion-v1-5" + sd_model = "Jiali/stable-diffusion-1.5" prompt = "desk" image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/desk_depth.png" input_channels = 3 out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_depth_sd15v2.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_keypose_sd_v14(self): - adapter_model = "TencentARC/t2iadapter_keypose_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "person" - image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/person_keypose.png" - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_keypose_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_openpose_sd_v14(self): - adapter_model = "TencentARC/t2iadapter_openpose_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "person" - image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/iron_man_pose.png" - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_openpose_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_seg_sd_v14(self): - adapter_model = "TencentARC/t2iadapter_seg_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "motorcycle" - image_url = ( - "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/motor.png" - ) - input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_seg_sd14v1.npy" + out_url = "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main/stable_diffusion_adapter/sd_adapter_v15_zoe_depth.npy" image = load_image(image_url) expected_out = load_numpy(out_url) @@ -797,11 +636,11 @@ def test_stable_diffusion_adapter_seg_sd_v14(self): def test_stable_diffusion_adapter_zoedepth_sd_v15(self): adapter_model = "TencentARC/t2iadapter_zoedepth_sd15v1" - sd_model = "runwayml/stable-diffusion-v1-5" + sd_model = "Jiali/stable-diffusion-1.5" prompt = "motorcycle" image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/motorcycle.png" input_channels = 3 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_zoedepth_sd15v1.npy" + out_url = "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main/stable_diffusion_adapter/sd_adapter_v15_zoe_depth.npy" image = load_image(image_url) expected_out = load_numpy(out_url) @@ -819,70 +658,13 @@ def test_stable_diffusion_adapter_zoedepth_sd_v15(self): max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) assert max_diff < 1e-2 - def test_stable_diffusion_adapter_canny_sd_v14(self): - adapter_model = "TencentARC/t2iadapter_canny_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "toy" - image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/toy_canny.png" - input_channels = 1 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_canny_sd14v1.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - def test_stable_diffusion_adapter_canny_sd_v15(self): adapter_model = "TencentARC/t2iadapter_canny_sd15v2" - sd_model = "runwayml/stable-diffusion-v1-5" + sd_model = "Jiali/stable-diffusion-1.5" prompt = "toy" image_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/toy_canny.png" input_channels = 1 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_canny_sd15v2.npy" - - image = load_image(image_url) - expected_out = load_numpy(out_url) - if input_channels == 1: - image = image.convert("L") - - adapter = T2IAdapter.from_pretrained(adapter_model, torch_dtype=torch.float16) - - pipe = StableDiffusionAdapterPipeline.from_pretrained(sd_model, adapter=adapter, safety_checker=None) - - pipe.to(torch_device) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing() - - generator = torch.Generator(device="cpu").manual_seed(0) - - out = pipe(prompt=prompt, image=image, generator=generator, num_inference_steps=2, output_type="np").images - - max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) - assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_sketch_sd14(self): - adapter_model = "TencentARC/t2iadapter_sketch_sd14v1" - sd_model = "CompVis/stable-diffusion-v1-4" - prompt = "cat" - image_url = ( - "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/edge.png" - ) - input_channels = 1 - out_url = "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/t2iadapter_sketch_sd14v1.npy" + out_url = "https://huggingface.co/datasets/diffusers/test-arrays/resolve/main/stable_diffusion_adapter/sd_adapter_v15_zoe_depth.npy" image = load_image(image_url) expected_out = load_numpy(out_url) @@ -906,7 +688,7 @@ def test_stable_diffusion_adapter_sketch_sd14(self): def test_stable_diffusion_adapter_sketch_sd15(self): adapter_model = "TencentARC/t2iadapter_sketch_sd15v2" - sd_model = "runwayml/stable-diffusion-v1-5" + sd_model = "Jiali/stable-diffusion-1.5" prompt = "cat" image_url = ( "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/edge.png" @@ -933,25 +715,3 @@ def test_stable_diffusion_adapter_sketch_sd15(self): max_diff = numpy_cosine_similarity_distance(out.flatten(), expected_out.flatten()) assert max_diff < 1e-2 - - def test_stable_diffusion_adapter_pipeline_with_sequential_cpu_offloading(self): - torch.cuda.empty_cache() - torch.cuda.reset_max_memory_allocated() - torch.cuda.reset_peak_memory_stats() - - adapter = T2IAdapter.from_pretrained("TencentARC/t2iadapter_seg_sd14v1") - pipe = StableDiffusionAdapterPipeline.from_pretrained( - "CompVis/stable-diffusion-v1-4", adapter=adapter, safety_checker=None - ) - pipe.set_progress_bar_config(disable=None) - pipe.enable_attention_slicing(1) - pipe.enable_sequential_cpu_offload() - - image = load_image( - "https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/t2i_adapter/motor.png" - ) - - pipe(prompt="foo", image=image, num_inference_steps=2) - - mem_bytes = torch.cuda.max_memory_allocated() - assert mem_bytes < 5 * 10**9 diff --git a/tests/pipelines/stable_diffusion_safe/test_safe_diffusion.py b/tests/pipelines/stable_diffusion_safe/test_safe_diffusion.py index 14100ea03dc1..cfaaa0914d17 100644 --- a/tests/pipelines/stable_diffusion_safe/test_safe_diffusion.py +++ b/tests/pipelines/stable_diffusion_safe/test_safe_diffusion.py @@ -277,7 +277,7 @@ def tearDown(self): torch.cuda.empty_cache() def test_harm_safe_stable_diffusion(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None) + sd_pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None) sd_pipe.scheduler = LMSDiscreteScheduler.from_config(sd_pipe.scheduler.config) sd_pipe = sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -338,7 +338,7 @@ def test_harm_safe_stable_diffusion(self): assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2 def test_nudity_safe_stable_diffusion(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5", safety_checker=None) + sd_pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5", safety_checker=None) sd_pipe.scheduler = LMSDiscreteScheduler.from_config(sd_pipe.scheduler.config) sd_pipe = sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) @@ -392,7 +392,7 @@ def test_nudity_safe_stable_diffusion(self): assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2 def test_nudity_safetychecker_safe_stable_diffusion(self): - sd_pipe = StableDiffusionPipeline.from_pretrained("runwayml/stable-diffusion-v1-5") + sd_pipe = StableDiffusionPipeline.from_pretrained("Jiali/stable-diffusion-1.5") sd_pipe = sd_pipe.to(torch_device) sd_pipe.set_progress_bar_config(disable=None) diff --git a/tests/pipelines/stable_diffusion_xl/test_stable_diffusion_xl_k_diffusion.py b/tests/pipelines/stable_diffusion_xl/test_stable_diffusion_xl_k_diffusion.py index 60207d9a0e76..94ee9f0facc8 100644 --- a/tests/pipelines/stable_diffusion_xl/test_stable_diffusion_xl_k_diffusion.py +++ b/tests/pipelines/stable_diffusion_xl/test_stable_diffusion_xl_k_diffusion.py @@ -58,7 +58,7 @@ def test_stable_diffusion_xl(self): [prompt], generator=generator, guidance_scale=9.0, - num_inference_steps=20, + num_inference_steps=2, height=512, width=512, output_type="np", @@ -69,9 +69,7 @@ def test_stable_diffusion_xl(self): image_slice = image[0, -3:, -3:, -1] assert image.shape == (1, 512, 512, 3) - expected_slice = np.array( - [0.79600024, 0.796546, 0.80682373, 0.79428387, 0.7905743, 0.8008807, 0.786183, 0.7835959, 0.797892] - ) + expected_slice = np.array([0.5420, 0.5038, 0.2439, 0.5371, 0.4660, 0.1906, 0.5221, 0.4290, 0.2566]) assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2 @@ -90,7 +88,7 @@ def test_stable_diffusion_karras_sigmas(self): [prompt], generator=generator, guidance_scale=7.5, - num_inference_steps=15, + num_inference_steps=2, output_type="np", use_karras_sigmas=True, height=512, @@ -102,9 +100,7 @@ def test_stable_diffusion_karras_sigmas(self): image_slice = image[0, -3:, -3:, -1] assert image.shape == (1, 512, 512, 3) - expected_slice = np.array( - [0.9506951, 0.9527786, 0.95309967, 0.9511477, 0.952523, 0.9515326, 0.9511933, 0.9480397, 0.94930184] - ) + expected_slice = np.array([0.6418, 0.6424, 0.6462, 0.6271, 0.6314, 0.6295, 0.6249, 0.6339, 0.6335]) assert np.abs(image_slice.flatten() - expected_slice).max() < 1e-2 @@ -124,7 +120,7 @@ def test_stable_diffusion_noise_sampler_seed(self): generator=torch.manual_seed(seed), noise_sampler_seed=seed, guidance_scale=9.0, - num_inference_steps=20, + num_inference_steps=2, output_type="np", height=512, width=512, @@ -134,7 +130,7 @@ def test_stable_diffusion_noise_sampler_seed(self): generator=torch.manual_seed(seed), noise_sampler_seed=seed, guidance_scale=9.0, - num_inference_steps=20, + num_inference_steps=2, output_type="np", height=512, width=512, diff --git a/tests/pipelines/test_pipelines_auto.py b/tests/pipelines/test_pipelines_auto.py index d060963f49d0..b899cf240d52 100644 --- a/tests/pipelines/test_pipelines_auto.py +++ b/tests/pipelines/test_pipelines_auto.py @@ -40,7 +40,7 @@ PRETRAINED_MODEL_REPO_MAPPING = OrderedDict( [ - ("stable-diffusion", "runwayml/stable-diffusion-v1-5"), + ("stable-diffusion", "Jiali/stable-diffusion-1.5"), ("if", "DeepFloyd/IF-I-XL-v1.0"), ("kandinsky", "kandinsky-community/kandinsky-2-1"), ("kandinsky22", "kandinsky-community/kandinsky-2-2-decoder"), @@ -539,7 +539,7 @@ def test_from_pipe_consistent(self): def test_controlnet(self): # test from_pretrained - model_repo = "runwayml/stable-diffusion-v1-5" + model_repo = "Jiali/stable-diffusion-1.5" controlnet_repo = "lllyasviel/sd-controlnet-canny" controlnet = ControlNetModel.from_pretrained(controlnet_repo, torch_dtype=torch.float16) diff --git a/tests/pipelines/text_to_video_synthesis/test_text_to_video_zero.py b/tests/pipelines/text_to_video_synthesis/test_text_to_video_zero.py index af59c7909a55..9038e3b0100f 100644 --- a/tests/pipelines/text_to_video_synthesis/test_text_to_video_zero.py +++ b/tests/pipelines/text_to_video_synthesis/test_text_to_video_zero.py @@ -40,7 +40,7 @@ def tearDown(self): torch.cuda.empty_cache() def test_full_model(self): - model_id = "runwayml/stable-diffusion-v1-5" + model_id = "Jiali/stable-diffusion-1.5" pipe = TextToVideoZeroPipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda") pipe.scheduler = DDIMScheduler.from_config(pipe.scheduler.config) generator = torch.Generator(device="cuda").manual_seed(0) diff --git a/tests/single_file/test_stable_diffusion_controlnet_img2img_single_file.py b/tests/single_file/test_stable_diffusion_controlnet_img2img_single_file.py index 3e4c1eaaa562..5127b9e745d8 100644 --- a/tests/single_file/test_stable_diffusion_controlnet_img2img_single_file.py +++ b/tests/single_file/test_stable_diffusion_controlnet_img2img_single_file.py @@ -30,11 +30,11 @@ @require_torch_gpu class StableDiffusionControlNetPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionControlNetPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_pruned.safetensors" + ckpt_path = "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors" original_config = ( "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml" ) - repo_id = "Lykon/dreamshaper-8" + repo_id = "Jiali/stable-diffusion-1.5" def setUp(self): super().setUp() diff --git a/tests/single_file/test_stable_diffusion_controlnet_inpaint_single_file.py b/tests/single_file/test_stable_diffusion_controlnet_inpaint_single_file.py index d7ccdbd89cc8..9d6576078a9c 100644 --- a/tests/single_file/test_stable_diffusion_controlnet_inpaint_single_file.py +++ b/tests/single_file/test_stable_diffusion_controlnet_inpaint_single_file.py @@ -29,9 +29,9 @@ @require_torch_gpu class StableDiffusionControlNetInpaintPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionControlNetInpaintPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_INPAINTING.inpainting.safetensors" + ckpt_path = "https://huggingface.co/botp/stable-diffusion-v1-5-inpainting/blob/main/sd-v1-5-inpainting.ckpt" original_config = "https://raw.githubusercontent.com/runwayml/stable-diffusion/main/configs/stable-diffusion/v1-inpainting-inference.yaml" - repo_id = "Lykon/dreamshaper-8-inpainting" + repo_id = "botp/stable-diffusion-v1-5-inpainting" def setUp(self): super().setUp() diff --git a/tests/single_file/test_stable_diffusion_controlnet_single_file.py b/tests/single_file/test_stable_diffusion_controlnet_single_file.py index 4bd7f025f64a..13d64dab77a1 100644 --- a/tests/single_file/test_stable_diffusion_controlnet_single_file.py +++ b/tests/single_file/test_stable_diffusion_controlnet_single_file.py @@ -29,11 +29,11 @@ @require_torch_gpu class StableDiffusionControlNetPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionControlNetPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_pruned.safetensors" + ckpt_path = "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors" original_config = ( "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml" ) - repo_id = "Lykon/dreamshaper-8" + repo_id = "Jiali/stable-diffusion-1.5" def setUp(self): super().setUp() diff --git a/tests/single_file/test_stable_diffusion_img2img_single_file.py b/tests/single_file/test_stable_diffusion_img2img_single_file.py index cbb5e9c3ee0e..fd99c4dede2f 100644 --- a/tests/single_file/test_stable_diffusion_img2img_single_file.py +++ b/tests/single_file/test_stable_diffusion_img2img_single_file.py @@ -23,11 +23,11 @@ @require_torch_gpu class StableDiffusionImg2ImgPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionImg2ImgPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_pruned.safetensors" + ckpt_path = "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors" original_config = ( "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml" ) - repo_id = "Lykon/dreamshaper-8" + repo_id = "Jiali/stable-diffusion-1.5" def setUp(self): super().setUp() diff --git a/tests/single_file/test_stable_diffusion_inpaint_single_file.py b/tests/single_file/test_stable_diffusion_inpaint_single_file.py index 3e133c6ea923..ba9583639b98 100644 --- a/tests/single_file/test_stable_diffusion_inpaint_single_file.py +++ b/tests/single_file/test_stable_diffusion_inpaint_single_file.py @@ -23,9 +23,9 @@ @require_torch_gpu class StableDiffusionInpaintPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionInpaintPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_INPAINTING.inpainting.safetensors" + ckpt_path = "https://huggingface.co/botp/stable-diffusion-v1-5-inpainting/blob/main/sd-v1-5-inpainting.ckpt" original_config = "https://raw.githubusercontent.com/runwayml/stable-diffusion/main/configs/stable-diffusion/v1-inpainting-inference.yaml" - repo_id = "Lykon/dreamshaper-8-inpainting" + repo_id = "botp/stable-diffusion-v1-5-inpainting" def setUp(self): super().setUp() @@ -63,7 +63,7 @@ def test_single_file_format_inference_is_same_as_pretrained(self): def test_single_file_loading_4_channel_unet(self): # Test loading single file inpaint with a 4 channel UNet - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_pruned.safetensors" + ckpt_path = "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors" pipe = self.pipeline_class.from_single_file(ckpt_path) assert pipe.unet.config.in_channels == 4 diff --git a/tests/single_file/test_stable_diffusion_single_file.py b/tests/single_file/test_stable_diffusion_single_file.py index 1283d4d99127..d40af28b2407 100644 --- a/tests/single_file/test_stable_diffusion_single_file.py +++ b/tests/single_file/test_stable_diffusion_single_file.py @@ -26,11 +26,11 @@ @require_torch_gpu class StableDiffusionPipelineSingleFileSlowTests(unittest.TestCase, SDSingleFileTesterMixin): pipeline_class = StableDiffusionPipeline - ckpt_path = "https://huggingface.co/Lykon/DreamShaper/blob/main/DreamShaper_8_pruned.safetensors" + ckpt_path = "https://huggingface.co/Jiali/stable-diffusion-1.5/blob/main/v1-5-pruned-emaonly.safetensors" original_config = ( "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/configs/stable-diffusion/v1-inference.yaml" ) - repo_id = "Lykon/dreamshaper-8" + repo_id = "Jiali/stable-diffusion-1.5" def setUp(self): super().setUp()