IndexError: index 29 is out of bounds for dimension 0 with size 29

Describe the bug

I have three problems because of the same reason.

1. TypeError: unsupported operand type(s) for +=: 'NoneType' and 'int'
   # upon completion increase step index by one
   self._step_index += 1 <---Error here
2. IndexError: index 29 is out of bounds for dimension 0 with size 29
   sigma_next = self.sigmas[self.step_index + 1] <--- Error here
3. RuntimeError: Already borrowed
   if _truncation is not None:
   self._tokenizer.no_truncation() <--- Error here
   Example: https://github.com/huggingface/tokenizers/issues/537
   The reason, as I understood, is threads. Do you know, how can I solve this problem?

Reproduction

from diffusers import (
    FluxPipeline,
    FlowMatchEulerDiscreteScheduler,
)
import torch

pipeline = FluxPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-schnell", torch_dtype=torch.bfloat16
).to("cuda")

seed = 42
height = 720
width = 1280

generator = torch.Generator(device="cuda").manual_seed(seed)

pipeline(
    prompt=prompt + ", highly detailed, all is depicted as silhouettes, without words",
    guidance_scale=0.,
    # num_inference_steps=10,
    height=height,
    width=width,
    generator=generator,
    max_sequence_length=256,
).images[0]

Logs

For example:
 Traceback (most recent call last):
   File "/opt/conda/lib/python3.10/site-packages/flask/app.py", line 1473, in wsgi_app
     response = self.full_dispatch_request()
   File "/opt/conda/lib/python3.10/site-packages/flask/app.py", line 882, in full_dispatch_request
     rv = self.handle_user_exception(e)
   File "/opt/conda/lib/python3.10/site-packages/flask/app.py", line 880, in full_dispatch_request
     rv = self.dispatch_request()
   File "/opt/conda/lib/python3.10/site-packages/flask/app.py", line 865, in dispatch_request
     return self.ensure_sync(self.view_functions[rule.endpoint])(**view_args)  # type: ignore[no-any-return]
   File "/app/main.py", line 29, in generate_image
     image = imagegen.run(**data)
   File "/app/image_generator.py", line 102, in run
     return generate_image()
   File "/app/image_generator.py", line 89, in generate_image
     return self.pipeline(
   File "/opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
     return func(*args, **kwargs)
   File "/opt/conda/lib/python3.10/site-packages/diffusers/pipelines/flux/pipeline_flux.py", line 734, in __call__
     latents = self.scheduler.step(noise_pred, t, latents, return_dict=False)[0]
  File "/opt/conda/lib/python3.10/site-packages/diffusers/schedulers/scheduling_flow_match_euler_discrete.py", line 295, in step
     sigma_next = self.sigmas[self.step_index + 1]
TypeError: unsupported operand type(s) for +: 'NoneType' and 'int'

System Info

• 🤗 Diffusers version: 0.31.0.dev0
• Platform: Linux-5.4.0-171-generic-x86_64-with-glibc2.35
• Running on Google Colab?: No
• Python version: 3.10.13
• PyTorch version (GPU?): 2.2.1 (True)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Huggingface_hub version: 0.24.6
• Transformers version: 4.44.2
• Accelerate version: 0.34.0
• PEFT version: 0.12.0
• Bitsandbytes version: not installed
• Safetensors version: 0.4.4
• xFormers version: not installed
• Accelerator: NVIDIA RTX A6000, 46068 MiB
• Using GPU in script?:
• Using distributed or parallel set-up in script?:

Who can help?

@yiyixuxu @sayakpaul @DN6

まず image_generator.py と pipeline_flux.py を通るスレッド化された Flask パスを再現し、次に scheduling_flow_match_euler_discrete.py の295-303行付近と tokenizer の切り詰め処理のパスを調査します。並行使用時に報告された3つの失敗を比較し、scheduler と tokenizer の状態がどのように共有されているかを追跡します。リクエストを並行して実行しても、提供された再現手順でこれらのエラーが発生しなくなれば完了です。