.. |
all_reduce
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
10 months ago |
attention
|
251568470e
initial nvidia fp8 e4m3 for kv cache
|
7 months ago |
backup
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
11 months ago |
cpu
|
b746fb5562
fix a few warnings on the cpu kernels
|
7 months ago |
hadamard
|
5d288aa76c
feat: add fast hadamard transformation kernels (#232)
|
1 year ago |
moe
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
10 months ago |
punica
|
d4edba99f9
add lora dims for Qwen1.5-32B
|
7 months ago |
quantization
|
2313c97e3d
add cutlass w8a8 kernels (#556)
|
7 months ago |
activation_kernels.cu
|
3d6695cfbb
feat: add approximate gelu activation kernels (#370)
|
11 months ago |
cache.h
|
251568470e
initial nvidia fp8 e4m3 for kv cache
|
7 months ago |
cache_kernels.cu
|
251568470e
initial nvidia fp8 e4m3 for kv cache
|
7 months ago |
cuda_compat.h
|
e3f2ea4850
make punica kernels work with rocm
|
7 months ago |
cuda_utils.h
|
31c95011a6
feat: FP8 E5M2 KV Cache (#226)
|
1 year ago |
cuda_utils_kernels.cu
|
31c95011a6
feat: FP8 E5M2 KV Cache (#226)
|
1 year ago |
dispatch_utils.h
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
11 months ago |
layernorm_kernels.cu
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
10 months ago |
ops.h
|
35ae01d7ba
refactor: attention metadata term
|
7 months ago |
pos_encoding_kernels.cu
|
e702f587cf
feat: add batched RoPE kernels (#371)
|
11 months ago |
pybind.cpp
|
2351a0e2cd
feat: FlashInfer backend for decoding phase (#548)
|
7 months ago |
reduction.cuh
|
9d81716bfd
[v0.5.3] Release Candidate (#388)
|
10 months ago |