[textual_inversion_sdxl.py] fix lr scheduler steps count

[yuanjua]

What does this PR do?

Part of #8384
Fix for LR in a distributed training when num_train_epoch is passed

run with the following scripts:

from huggingface_hub import snapshot_download

local_dir = "./cat"
snapshot_download(
    "diffusers/cat_toy_example", local_dir=local_dir, repo_type="dataset", ignore_patterns=".gitattributes"
)

#!/bin/bash

mkdir -p textual_inversion_cat/logs

accelerate launch \
  --multi_gpu \
  --num_processes=2 \
  --mixed_precision="bf16" \
  textual_inversion_sdxl.py \
  --pretrained_model_name_or_path="stabilityai/stable-diffusion-xl-base-1.0" \
  --train_data_dir="./cat" \
  --output_dir="textual_inversion_cat" \
  --placeholder_token="<my-token>" \
  --initializer_token="cat" \
  --resolution=1024 \
  --train_batch_size=1 \
  --gradient_accumulation_steps=1 \
  --max_train_steps=100 \
  --learning_rate=1e-4 \
  --num_vectors=4 \
  --validation_prompt="a photo of a <my-token>" \
  --num_validation_images=4 \
  --validation_steps=100 \
  --save_steps=500 \
  --mixed_precision="bf16" \
  --enable_xformers_memory_efficient_attention \
  --gradient_checkpointing \
  --center_crop \
  --report_to="tensorboard" \
  --logging_dir="textual_inversion_cat/logs"

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• Did you read our philosophy doc (important for complex PRs)?
• Was this discussed/approved via a GitHub issue or the forum? Please add a link to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

@sayakpaul

[sayakpaul]

Thank you.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.