Support schnell lora training

Hi @kohya-ss, I just find the current training script or code cannot be applied to FLUX schnell directly. According to @sdbds, it seems the shift problem? There are a lot of users prefer schnell, can you help to launch a schnell version for lora training ? that will be very helpful !!!!!!
I change the script to the following, but it still gets bad result.
```
--timestep_sampling uniform \
--discrete_flow_shift 1.0 \
--guidance_scale 0.0 \
```

The issue does not name a file, entry point, or test. Start by locating the existing FLUX LoRA training path and the handling of --timestep_sampling, --discrete_flow_shift, and --guidance_scale; compare it with the FLUX schnell requirements described in the issue. Done means schnell training produces acceptable LoRA results and is covered by an appropriate validation or regression test.

Written by the indexing model from the issue text.