add option to return images by cycle step

[chrisgoringe]

This PR adds a boolean input to the detailer nodes, return_images_by_cycle_step.

Default value is False, leaving behaviour unchanged.

If set to True, the image after each cycle step is returned in a batch.

This allows the user to specify a number of cycles, and then pick the best result.

(I use this in combination with https://github.com/chrisgoringe/cg-image-filter which allows me to pick the best image to continue with)

[ltdrdata]

It crashes when return_by_cycle_step is disabled.

: 640x384 1 face, 5.0ms
Speed: 0.8ms preprocess, 5.0ms inference, 2.9ms postprocess per image at shape (1, 3, 640, 384)
Detailer: segment upscale for ((np.float32(66.21362), np.float32(84.172455))) | crop region (198, 252) x 3.047865629196167 -> (603, 768)
 40%|██████████████████████████████████████████████████████████████████████████████████████████████████▍                                                                                                                                                   | 8/20 [00:00<00:00, 12.41it/s]FETCH ComfyRegistry Data: 75/107
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 20/20 [00:01<00:00, 12.26it/s]
[Impact Pack] vae decoded in 0.1s
!!! Exception during processing !!! cannot access local variable 'refined_images_by_step' where it is not associated with a value
Traceback (most recent call last):
  File "/mnt/teratera/git/ComfyUI/execution.py", line 510, in execute
    output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, hidden_inputs=hidden_inputs)
                                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/teratera/git/ComfyUI/execution.py", line 324, in get_output_data
    return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, hidden_inputs=hidden_inputs)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/teratera/git/ComfyUI/execution.py", line 298, in _async_map_node_over_list
    await process_inputs(input_dict, i)
  File "/mnt/teratera/git/ComfyUI/execution.py", line 286, in process_inputs
    result = f(**inputs)
             ^^^^^^^^^^^
  File "/mnt/teratera/git/ComfyUI/custom_nodes/comfyui-impact-pack/modules/impact/impact_pack.py", line 899, in doit
    enhanced_img, cropped_enhanced, cropped_enhanced_alpha, mask, cnet_pil_list = FaceDetailer.enhance_face(
                                                                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/teratera/git/ComfyUI/custom_nodes/comfyui-impact-pack/modules/impact/impact_pack.py", line 853, in enhance_face
    DetailerForEach.do_detail(image, segs, model, clip, vae, guide_size, guide_size_for_bbox, max_size, seed, steps, cfg,
  File "/mnt/teratera/git/ComfyUI/custom_nodes/comfyui-impact-pack/modules/impact/impact_pack.py", line 365, in do_detail
    result = core.enhance_detail(cropped_image, model, clip, vae, guide_size, guide_size_for_bbox, max_size,
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/teratera/git/ComfyUI/custom_nodes/comfyui-impact-pack/modules/impact/core.py", line 431, in enhance_detail
    refined_images_by_step = [ utils.tensor_resize(i,w,h) for i in refined_images_by_step ]
                                                                   ^^^^^^^^^^^^^^^^^^^^^^
UnboundLocalError: cannot access local variable 'refined_images_by_step' where it is not associated with a value