wangshuai09
4e2d95e372
[Hardware][ROCM] using current_platform.is_rocm ( #9642 )
...
Signed-off-by: wangshuai09 <391746016@qq.com>
2024-10-28 04:07:00 +00:00
Charlie Fu
59449095ab
[Performance][Kernel] Fused_moe Performance Improvement ( #9384 )
...
Signed-off-by: charlifu <charlifu@amd.com>
2024-10-24 15:37:52 -07:00
bnellnm
eca2c5f7c0
[Bugfix] Fix support for dimension like integers and ScalarType ( #9299 )
2024-10-17 19:08:34 +00:00
ElizaWszola
05d686432f
[Kernel] Zero point support in fused MarlinMoE kernel + AWQ Fused MoE ( #8973 )
...
Co-authored-by: Dipika <dipikasikka1@gmail.com>
Co-authored-by: Dipika Sikka <ds3822@columbia.edu>
2024-10-04 12:34:44 -06:00
ElizaWszola
d081da0064
[Bugfix] Fix Marlin MoE act order when is_k_full == False ( #8741 )
...
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
2024-09-28 18:19:40 -07:00
bnellnm
300da09177
[Kernel] Fullgraph and opcheck tests ( #8479 )
2024-09-25 08:35:52 -06:00
Cyrus Leung
6ffa3f314c
[CI/Build] Avoid CUDA initialization ( #8534 )
2024-09-18 10:38:11 +00:00
ElizaWszola
a091e2da3e
[Kernel] Enable 8-bit weights in Fused Marlin MoE ( #8032 )
...
Co-authored-by: Dipika <dipikasikka1@gmail.com>
2024-09-16 09:47:19 -06:00
Dipika Sikka
6cd5e5b07e
[Misc] Fused MoE Marlin support for GPTQ ( #8217 )
2024-09-09 23:02:52 -04:00
jon-chuang
50b8d08dbd
[Misc/Testing] Use torch.testing.assert_close ( #7324 )
2024-08-16 04:24:04 +00:00
Robert Shaw
7c008c51a9
[ Misc ] Refactor MoE to isolate Fp8 From Mixtral ( #5970 )
...
Co-authored-by: Robert Shaw <rshaw@neuralmagic>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
2024-07-02 21:54:35 +00:00
Avshalom Manevich
12a59959ed
[Bugfix] adding chunking mechanism to fused_moe to handle large inputs ( #6029 )
2024-07-01 21:08:29 +00:00
Michael Goin
2a052011ca
[Kernel] Support MoE Fp8 Checkpoints for Mixtral (Static Weights with Dynamic/Static Activations) ( #4527 )
...
Follow on to #4332 to enable FP8 checkpoint loading for Mixtral and supersedes #4436 .
This PR enables the following checkpoint loading features for Mixtral:
Supports loading fp8 checkpoints for Mixtral, such as this "nm-testing/Mixtral-8x7B-Instruct-v0.1-FP8" test model
Supports static or dynamic activation quantization with static weight quantization (all per tensor)
Supports different scales for each expert weight
Supports Fp8 in QKV layer
Notes:
The Expert Gate/Router always runs at half / full precision for now.
If there are different weight scales between QKV layer (for separate QKV weights), they are re-quantized using layer.weight_scale.max() so we can have a single gemm for performance.
2024-05-04 11:45:16 -07:00
Antoni Baum
a10d3056da
[Core] Set linear_weights directly on the layer ( #3977 )
2024-04-11 16:35:51 -04:00
SangBin Cho
01bfb22b41
[CI] Try introducing isort. ( #3495 )
2024-03-25 07:59:47 -07:00
Nick Hill
41deac4a3d
[BugFix] 1D query fix for MoE models ( #3597 )
2024-03-24 16:00:16 -07:00
Zhuohan Li
2f8844ba08
Re-enable the 80 char line width limit ( #3305 )
2024-03-10 19:49:14 -07:00
Woosuk Kwon
f0d4e14557
Add fused top-K softmax kernel for MoE ( #2769 )
2024-02-05 17:38:02 -08:00
Philipp Moritz
d0d93b92b1
Add unit test for Mixtral MoE layer ( #2677 )
2024-01-31 14:34:17 -08:00
