Add torch_xla and from_single_file to instruct-pix2pix

[hlky]

What does this PR do?

Replaces #10355

from_single_file changes were determined by comparing v1-5-pruned.safetensors to instruct-pix2pix-00-22000.safetensors

import safetensors
from huggingface_hub import hf_hub_download

pix2pix = safetensors.safe_open(hf_hub_download("timbrooks/instruct-pix2pix", filename="instruct-pix2pix-00-22000.safetensors"), framework="pt")

v1 = safetensors.safe_open(hf_hub_download("stable-diffusion-v1-5/stable-diffusion-v1-5", filename="v1-5-pruned.safetensors"), framework="pt")

extra_keys = set(pix2pix.keys()) - set(v1.keys())
print(extra_keys)

for key in pix2pix.keys():
    tensor = pix2pix.get_tensor(key)
    tensor_v1 = v1.get_tensor(key)
    if tensor.shape != tensor_v1.shape:
        print(f"{key} {tensor.shape} (pix2pix) != {tensor_v1.shape} (v1)")

set()
model.diffusion_model.input_blocks.0.0.weight torch.Size([320, 8, 3, 3]) (pix2pix) != torch.Size([320, 4, 3, 3]) (v1)
model_ema.diffusion_modelinput_blocks00weight torch.Size([320, 8, 3, 3]) (pix2pix) != torch.Size([320, 4, 3, 3]) (v1)

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@sayakpaul @chaowenguo

[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.

[sayakpaul]

Can we also have a nightly test for this?