Address compatibility issues arising from the removal of the XLA dependency

[weiyu0824]

Description

This PR addresses compatibility issues arising from the removal of the XLA dependency in upstream vLLM (reference.

Changes:

• Moved PallasAttention static function from vllm to tpu-inference.
• Update TpuPlatform in tpu_platform.py to use local PallasAttentionBackend.
• Add create_weights in compressed_tensors_w8a8_int8.py because vllm's implementation would introduce another error.
• Register PallasAttention using decorator (see line:32 and line:37 in tpu_inference/layers/vllm/attention.py) because we already remove Pallas enum in upstream.

Tests

• buildkite result

Checklist

Before submitting this PR, please make sure:

• I have performed a self-review of my code.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have made or will make corresponding changes to any relevant documentation.

[github-actions]

Description

Start with a short description of what the PR does and how this is a change from
the past.

The rest of the description includes relevant details and context, examples:

• why is this change being made,
• the problem being solved and any relevant context,
• why this is a good solution,
• some information about the specific implementation,
• shortcomings of the solution and possible future improvements.

If the change fixes a Github issue, please include a link, e.g.,:
FIXES: #123456

Tests

Please describe how you tested this change, and include any instructions and/or
commands to reproduce.

Checklist

Before submitting this PR, please make sure:

• I have performed a self-review of my code.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have made or will make corresponding changes to any relevant documentation.

[QiliangCui]

can you explain more about why

Add create_weights in compressed_tensors_w8a8_int8.py because vllm's implementation would introduce another error.

[kyuyeunk]

seems like ci is failing due to upstream change: vllm-project/vllm#30519

I think this should be a simple fix. @weiyu0824 can you add a fix commit to this branch as well?

[kyuyeunk]

hmmm we have another error due to upstream change 1 hour ago: vllm-project/vllm#32003

yeah, feel free to ignore all my other comments and let's just try to get this pr merged asap.