6676 port diffusion schedulers (#7364)

This is an update to PR #7332
- I addressed the comments but failed to push the changes before it was
merged! Changes are very minor.

### Description

A few sentences describing the changes proposed in this pull request.

### Types of changes
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [ ] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

---------

Signed-off-by: Mark Graham <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

monai/networks/schedulers/ddim.py

@@ -34,23 +34,10 @@
 import numpy as np
 import torch
 
-from monai.utils import StrEnum
-
+from .ddpm import DDPMPredictionType
 from .scheduler import Scheduler
 
-
-class DDIMPredictionType(StrEnum):
-    """
-    Set of valid prediction type names for the DDIM scheduler's `prediction_type` argument.
-
-    epsilon: predicting the noise of the diffusion process
-    sample: directly predicting the noisy sample
-    v_prediction: velocity prediction, see section 2.4 https://imagen.research.google/video/paper.pdf
-    """
-
-    EPSILON = "epsilon"
-    SAMPLE = "sample"
-    V_PREDICTION = "v_prediction"
+DDIMPredictionType = DDPMPredictionType
 
 
 class DDIMScheduler(Scheduler):
@@ -126,6 +113,13 @@ def set_timesteps(self, num_inference_steps: int, device: str | torch.device | N
 
         self.num_inference_steps = num_inference_steps
         step_ratio = self.num_train_timesteps // self.num_inference_steps
+        if self.steps_offset >= step_ratio:
+            raise ValueError(
+                f"`steps_offset`: {self.steps_offset} cannot be greater than or equal to "
+                f"`num_train_timesteps // num_inference_steps : {step_ratio}` as this will cause timesteps to exceed"
+                f" the max train timestep."
+            )
+
         # creates integer timesteps by multiplying by ratio
         # casting to int to avoid issues when num_inference_step is power of 3
         timesteps = (np.arange(0, num_inference_steps) * step_ratio).round()[::-1].copy().astype(np.int64)
@@ -159,7 +153,6 @@ def step(
             timestep: current discrete timestep in the diffusion chain.
             sample: current instance of sample being created by diffusion process.
             eta: weight of noise for added noise in diffusion step.
-            predict_epsilon: flag to use when model predicts the samples directly instead of the noise, epsilon.
             generator: random number generator.
 
         Returns: