and bump doc version.

This PR should fix #330 .

when `128 / page == 0`, our binary search might run into division by
zero issue.

For small query size setting, we might use large kv tile size.

…gedKVCache` class (#337)

Add more tests for coverage.

This PR supports customized logits soft cap values. Different models
might use different logits soft cap values (e.g. Grok-1 uses 30 and
Gemma-2 uses 50).

🤖 I have created a release *beep* *boop*
---


##
[0.0.7](v0.0.6...v0.0.7)
(2024-06-28)

### Bugfix

* fix the `forward_return_lse` function in
`BatchPrefillWithRaggedKVCache` class
([#337](#337))
* fix the scheduler behavior of large page size
([#333](#333))

### Features

* customize `logits_soft_cap` value
([#339](#339))
([a2498f5](a2498f5))


### Performance Improvements

* change minimal `kv_chunk_size` back to 128
([#329](#329))
([f237f5f](f237f5f))
* more options for kv tile size
([#336](#336))
([bf2a6c7](bf2a6c7))

---
This PR was generated with [Release
Please](https://github.com/googleapis/release-please). See
[documentation](https://github.com/googleapis/release-please#release-please).

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Zihao Ye <expye@outlook.com>

This commit resolves a build-time error with the following message:

```
CMake Error at 3rdparty/flashinfer/CMakeLists.txt:313 (add_library):
  No SOURCES given to target: prefill_kernels
```

This occurred after
#266, which replaces the
`FLASHINFER_GEN_CASUALS` option with `FLASHINFER_GEN_MASK_MODES`.
However, the definition of `flashinfer_option(FLASHINFER_GEN_CASUALS ...
)` was not replaced. As a result, loop over the empty `MASK_MODES` does
not produce any kernels that should be compiled.

This commit updates the `flashinfer_option(FLASH_GEN_CASUALS ...)` line
to instead define `FLASH_GEN_MASK_MODES`, using the same default value
as `config.cmake`.

…r large wheels (#341)

Replace ssh with https

This PR refactors the batch decode related kernels, and make the
following breaking changes:
1. remove `batch_decode_with_padded_kv_cache` operator, we encourage
user to use `BatchDecodeWithPagedKVCacheWrapper`.
2. Delete redundant DTypeQ * DTypeKV combinations, now we only support
the following cases:
  1. DTypeQ == DTypeKV
  2. DTypeQ is a float16 and DTypeKV is a float8

The output data type follows the query data type.

Also reduce binary size but limit the maximum number of registers for
`x_frag` and `o_frag` to 200.

Do not compile `NUM_FRAGS_Z=6` to reduce wheel size.
Also revert #341 as they don't make effect.