FLUX Kontext. Transformers error: safetensors_rust.SafetensorError: Error while deserializing header: MetadataIncompleteBuffer

[tin2tin]

Describe the bug

Kontext worked yesterday, but today there is an error running the included example:
safetensors_rust.SafetensorError: Error while deserializing header: MetadataIncompleteBuffer

Reproduction

import torch
from diffusers import FluxKontextPipeline
from diffusers.utils import load_image

pipe = FluxKontextPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-Kontext-dev", torch_dtype=torch.bfloat16
)
pipe.to("cuda")

image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/yarn-art-pikachu.png").convert("RGB")
prompt = "Make Pikachu hold a sign that says 'Black Forest Labs is awesome', yarn art style, detailed, vibrant colors"
image = pipe(
    image=image,
    prompt=prompt,
    guidance_scale=2.5,
    generator=torch.Generator().manual_seed(42),
).images[0]
image.save("flux-kontext.png")

Logs

Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████| 3/3 [00:00<00:00, 68.34it/s]
Loading checkpoint shards:   0%|                                                                 | 0/2 [00:00<?, ?it/s]
Loading pipeline components...:  71%|█████████████████████████████████████▏              | 5/7 [00:01<00:00,  3.09it/s]
Error: Python: Traceback (most recent call last):
  File ".\python\Lib\site-packages\transformers\modeling_utils.py", line 311, in _wrapper
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File ".\python\Lib\site-packages\transformers\modeling_utils.py", line 4833, in from_pretrained
    ) = cls._load_pretrained_model(
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".\python\Lib\site-packages\transformers\modeling_utils.py", line 5296, in _load_pretrained_model
    _error_msgs, disk_offload_index, cpu_offload_index = load_shard_file(args)
                                                         ^^^^^^^^^^^^^^^^^^^^^
  File ".\python\Lib\site-packages\transformers\modeling_utils.py", line 920, in load_shard_file
    state_dict = load_state_dict(
                 ^^^^^^^^^^^^^^^^
  File ".\python\Lib\site-packages\transformers\modeling_utils.py", line 532, in load_state_dict
    with safe_open(checkpoint_file, framework="pt") as f:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
safetensors_rust.SafetensorError: Error while deserializing header: MetadataIncompleteBuffer

System Info

The latest alpha.

Who can help?

@DN6 @a-r-r-o-w

[a-r-r-o-w]

This does not seem like a diffusers side error to me. From some quick searches, it looks like this happens when a model is corrupted during download or later. Could you try deleting any existing checkpoints you have of this model and then downloading it again? The error hints to the text encoder weights being corrupted

[tin2tin]

Yes, thank you! Deleting the weights, and redownloading them, worked.