diffusers/tests/schedulers/test_scheduler_ddim_inverse.py at main · huggingface/diffusers · GitHub

Name: diffusers/tests/schedulers/test_scheduler_ddim_inverse.py at main · huggingface/diffusers · GitHub
Rating: 4.6 (3314 reviews)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
importunittest

importtorch

fromdiffusersimportDDIMInverseScheduler

from .test_schedulersimportSchedulerCommonTest


classDDIMInverseSchedulerTest(SchedulerCommonTest):
scheduler_classes= (DDIMInverseScheduler,)
forward_default_kwargs= (("num_inference_steps", 50),)

defget_scheduler_config(self, **kwargs):
config= {
"num_train_timesteps": 1000,
"beta_start": 0.0001,
"beta_end": 0.02,
"beta_schedule": "linear",
"clip_sample": True,
 }

config.update(**kwargs)
returnconfig

deffull_loop(self, **config):
scheduler_class=self.scheduler_classes[0]
scheduler_config=self.get_scheduler_config(**config)
scheduler=scheduler_class(**scheduler_config)

num_inference_steps=10

model=self.dummy_model()
sample=self.dummy_sample_deter

scheduler.set_timesteps(num_inference_steps)

fortinscheduler.timesteps:
residual=model(sample, t)
sample=scheduler.step(residual, t, sample).prev_sample

returnsample

deftest_timesteps(self):
fortimestepsin [100, 500, 1000]:
self.check_over_configs(num_train_timesteps=timesteps)

deftest_steps_offset(self):
forsteps_offsetin [0, 1]:
self.check_over_configs(steps_offset=steps_offset)

scheduler_class=self.scheduler_classes[0]
scheduler_config=self.get_scheduler_config(steps_offset=1)
scheduler=scheduler_class(**scheduler_config)
scheduler.set_timesteps(5)
asserttorch.equal(scheduler.timesteps, torch.LongTensor([1, 201, 401, 601, 801]))

deftest_betas(self):
forbeta_start, beta_endinzip([0.0001, 0.001, 0.01, 0.1], [0.002, 0.02, 0.2, 2]):
self.check_over_configs(beta_start=beta_start, beta_end=beta_end)

deftest_schedules(self):
forschedulein ["linear", "squaredcos_cap_v2"]:
self.check_over_configs(beta_schedule=schedule)

deftest_prediction_type(self):
forprediction_typein ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=prediction_type)

deftest_clip_sample(self):
forclip_samplein [True, False]:
self.check_over_configs(clip_sample=clip_sample)

deftest_timestep_spacing(self):
fortimestep_spacingin ["trailing", "leading"]:
self.check_over_configs(timestep_spacing=timestep_spacing)

deftest_rescale_betas_zero_snr(self):
forrescale_betas_zero_snrin [True, False]:
self.check_over_configs(rescale_betas_zero_snr=rescale_betas_zero_snr)

deftest_thresholding(self):
self.check_over_configs(thresholding=False)
forthresholdin [0.5, 1.0, 2.0]:
forprediction_typein ["epsilon", "v_prediction"]:
self.check_over_configs(
thresholding=True,
prediction_type=prediction_type,
sample_max_value=threshold,
 )

deftest_time_indices(self):
fortin [1, 10, 49]:
self.check_over_forward(time_step=t)

deftest_inference_steps(self):
fort, num_inference_stepsinzip([1, 10, 50], [10, 50, 500]):
self.check_over_forward(time_step=t, num_inference_steps=num_inference_steps)

@unittest.skip("Test not supported.")
deftest_add_noise_device(self):
pass

deftest_full_loop_no_noise(self):
sample=self.full_loop()

result_sum=torch.sum(torch.abs(sample))
result_mean=torch.mean(torch.abs(sample))

assertabs(result_sum.item() -671.6816) <1e-2
assertabs(result_mean.item() -0.8746) <1e-3

deftest_full_loop_with_v_prediction(self):
sample=self.full_loop(prediction_type="v_prediction")

result_sum=torch.sum(torch.abs(sample))
result_mean=torch.mean(torch.abs(sample))

assertabs(result_sum.item() -1394.2185) <1e-2
assertabs(result_mean.item() -1.8154) <1e-3

deftest_full_loop_with_set_alpha_to_one(self):
# We specify different beta, so that the first alpha is 0.99
sample=self.full_loop(set_alpha_to_one=True, beta_start=0.01)
result_sum=torch.sum(torch.abs(sample))
result_mean=torch.mean(torch.abs(sample))

assertabs(result_sum.item() -539.9622) <1e-2
assertabs(result_mean.item() -0.7031) <1e-3

deftest_full_loop_with_no_set_alpha_to_one(self):
# We specify different beta, so that the first alpha is 0.99
sample=self.full_loop(set_alpha_to_one=False, beta_start=0.01)
result_sum=torch.sum(torch.abs(sample))
result_mean=torch.mean(torch.abs(sample))

assertabs(result_sum.item() -542.6722) <1e-2
assertabs(result_mean.item() -0.7066) <1e-3