[LLM INFER] not use gemm_dequant default and fix bug

[yuanlehome]

PR types

Bug fixes

PR changes

Others

Description

not use gemm_dequant default and fix bug

bug error:

line 265, in get_block_shape_and_split_kv_block
    outs = _C_ops._run_custom_op("get_block_shape_and_split_kv_block", seq_lens_encoder,seq_lens_decoder,max_enc_len_this_time,seq_lens_this_time,cum_offsets,encoder_block_shape_q,decoder_block_shape_q,group_size,block_size,decoder_step_token_num)
TypeError: (InvalidType) argument (position 11) must be int, but got tuple (at ../paddle/fluid/pybind/eager_utils.cc:185)

[paddle-bot]

Thanks for your contribution!

[CLAassistant]

All committers have signed the CLA.

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 7 lines in your changes missing coverage. Please review.

Project coverage is 52.95%. Comparing base (0b4b810) to head (8fad937).
Report is 2 commits behind head on develop.

Files with missing lines	Patch %	Lines
...erimental/transformers/fused_transformer_layers.py	0.00%	4 Missing ⚠️
...lp/experimental/transformers/qwen2_moe/modeling.py	0.00%	3 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #9498      +/-   ##
===========================================
+ Coverage    52.84%   52.95%   +0.11%     
===========================================
  Files          688      688              
  Lines       109378   109381       +3     
===========================================
+ Hits         57797    57922     +125     
+ Misses       51581    51459     -122     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.