xiaoxiaohehe001

[Cherry-Pick] fix nvcc_lazy

1

### PR types Others ### PR changes Others ### Describe fix nvcc_lazy

[Paddle inference] Support nvcc_lazy.

1

### PR types Others ### PR changes Others ### Describe Support nvcc_lazy.

[LLM] Support pre_caches input of llama

5

### PR types Function optimization ### PR changes Others ### Description Support pre_caches input of llama

## Description When the input of Range is slice dynamic shapetensor, there is a error about rank == -1    Range: Before addFill, the rank of numberOfElements is...

support_qwen_infer

3

### PR types Others ### PR changes Others ### Description support_qwen

[LLM] Support QWEN BlockAttention && PTQ

4

### PR types New features ### PR changes Models ### Description Support QWEN BlockAttention && PTQ

[LLM] Support baichuan2-13b paddle52.

5

### PR types New features ### PR changes Models ### Description Support baichuan2-13b paddle52.

[LLM] support bloom fine grained dybatch v1.

7

### PR types New features ### PR changes Models ### Description support bloom fine grained dybatch v1.

[LLM] Support llama precache input.

5

### PR types New features ### PR changes Others ### Description Support llama precache input.

topptopk

3

### PR types ### PR changes ### Description