Commits · afd0da2186c1d58fb48e138df0a2f548612b5d7d · OpenDAS / vllm_cscc

31 Jan, 2025 2 commits
- [Kernel][Quantization] Integrate block-quantized CUTLASS kernels for DeepSeekV3 (#12587) · eb5741ad
  Tyler Michael Smith authored Jan 31, 2025
```
Integrates the block-quantized kernels introduced in
https://github.com/vllm-project/vllm/pull/11868

 for use in linear
layers.
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  eb5741ad
- [BugFix] Fix Torch.Compile For DeepSeek (#12594) · 325f679f
  Robert Shaw authored Jan 31, 2025
```
Co-authored-by: simon-mo <xmo@berkeley.edu>
```
  325f679f
23 Jan, 2025 1 commit

[BugFix] Fix parameter names and `process_after_weight_loading` for W4A16 MoE... · eb5cb5e5

Dipika Sikka authored Jan 23, 2025


[BugFix] Fix parameter names and `process_after_weight_loading` for W4A16 MoE Group Act Order  (#11528)
Signed-off-by: ElizaWszola <eliza@neuralmagic.com>
Co-authored-by: ElizaWszola <eliza@neuralmagic.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>

eb5cb5e5

17 Jan, 2025 1 commit
- [AMD][FP8] Using MI300 FP8 format on ROCm for block_quant (#12134) · b5b57e30
  Gregory Shtrasberg authored Jan 17, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
```
  b5b57e30
16 Jan, 2025 1 commit

[Core] Default to using per_token quantization for fp8 when cutlass is supported. (#8651) · fa0050db

Elfie Guo authored Jan 15, 2025


Signed-off-by: mgoin <michael@neuralmagic.com>
Co-authored-by: Michael Goin <mgoin@redhat.com>
Co-authored-by: mgoin <michael@neuralmagic.com>

fa0050db

09 Jan, 2025 1 commit

[Misc] Move `print_*_once` from utils to logger (#11298) · d848800e

Cyrus Leung authored Jan 09, 2025

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: Maxime Fournioux <55544262+mfournioux@users.noreply.github.com>
Co-authored-by: Maxime Fournioux <55544262+mfournioux@users.noreply.github.com>

d848800e

30 Dec, 2024 1 commit
- [CI/Build][CPU] Fix CPU CI by lazy importing triton FP8 kernels (#11618) · 5dbf8545
  Li, Jiang authored Dec 30, 2024
```
Signed-off-by: jiang1.li <jiang1.li@intel.com>
```
  5dbf8545
27 Dec, 2024 2 commits
- [BugFix] Fix quantization for all other methods (#11547) · 2339d59f
  Robert Shaw authored Dec 27, 2024
  
  2339d59f
- Deepseek v3 (#11502) · f49777ba
  Simon Mo authored Dec 26, 2024
```
Signed-off-by: mgoin <michael@neuralmagic.com>
Co-authored-by: mgoin <michael@neuralmagic.com>
Co-authored-by: robertgshaw2-neuralmagic <rshaw@neuralmagic.com>
```
  f49777ba
26 Dec, 2024 1 commit

[Model] [Quantization] Support deepseek_v3 w8a8 fp8 block-wise quantization (#11523) · 2072924d

Michael Goin authored Dec 26, 2024


Signed-off-by: mgoin <michael@neuralmagic.com>
Signed-off-by: simon-mo <simon.mo@hey.com>
Signed-off-by: simon-mo <xmo@berkeley.edu>
Co-authored-by: simon-mo <simon.mo@hey.com>
Co-authored-by: simon-mo <xmo@berkeley.edu>
Co-authored-by: HandH1998 <1335248067@qq.com>

2072924d

06 Nov, 2024 1 commit
- Remove ScaledActivation for AWQ (#10057) · 399c7986
  Michael Goin authored Nov 06, 2024
```
Signed-off-by: mgoin <michael@neuralmagic.com>
```
  399c7986
28 Oct, 2024 1 commit
- [Hardware][ROCM] using current_platform.is_rocm (#9642) · 4e2d95e3
  wangshuai09 authored Oct 28, 2024
```
Signed-off-by: wangshuai09 <391746016@qq.com>
```
  4e2d95e3
18 Sep, 2024 1 commit
- [CI/Build] Avoid CUDA initialization (#8534) · 6ffa3f31
  Cyrus Leung authored Sep 18, 2024
  
  6ffa3f31
30 Aug, 2024 1 commit

[Model] Adding support for MSFT Phi-3.5-MoE (#7729) · 1248e850

Wenxiang authored Aug 31, 2024


Co-authored-by: Your Name <you@example.com>
Co-authored-by: Zeqi Lin <zelin@microsoft.com>
Co-authored-by: Zeqi Lin <Zeqi.Lin@microsoft.com>

1248e850

27 Aug, 2024 1 commit
- [Kernel] Expand MoE weight loading + Add Fused Marlin MoE Kernel (#7766) · fc911880
  Dipika Sikka authored Aug 27, 2024
```
Co-authored-by: ElizaWszola <eliza@neuralmagic.com>
```
  fc911880
22 Aug, 2024 2 commits
- [Misc] update fp8 to use `vLLMParameter` (#7437) · 955b5191
  Dipika Sikka authored Aug 22, 2024
  
  955b5191
- Revert "[Kernel] Expand MoE weight loading + Add Fused Marlin MoE Kernel (#7527)" (#7764) · aae74ef9
  Michael Goin authored Aug 21, 2024
  
  aae74ef9
21 Aug, 2024 1 commit
- [Kernel] Expand MoE weight loading + Add Fused Marlin MoE Kernel (#7527) · 8678a69a
  Dipika Sikka authored Aug 21, 2024
```
Co-authored-by: ElizaWszola <eliza@neuralmagic.com>
```
  8678a69a
16 Aug, 2024 2 commits
- [Kernel] W8A16 Int8 inside FusedMoE (#7415) · 7fc23be8
  Mor Zusman authored Aug 16, 2024
  
  7fc23be8
- [Feature][Hardware][Amd] Add fp8 Linear Layer for Rocm (#7210) · e837b624
  Charlie Fu authored Aug 16, 2024
  
  e837b624
13 Aug, 2024 1 commit
- [Misc] Update Fused MoE weight loading (#7334) · d3bdfd3a
  Dipika Sikka authored Aug 13, 2024
  
  d3bdfd3a
07 Aug, 2024 1 commit
- [Bugfix][FP8] Fix dynamic FP8 Marlin quantization (#7219) · 5223199e
  Michael Goin authored Aug 07, 2024
  
  5223199e
29 Jul, 2024 1 commit
- [Bugfix] Allow vllm to still work if triton is not installed. (#6786) · 9a7e2d05
  Thomas Parnell authored Jul 29, 2024
```
Signed-off-by: Thomas Parnell <tpa@zurich.ibm.com>
```
  9a7e2d05
25 Jul, 2024 1 commit
- [ Misc ] `fp8-marlin` channelwise via `compressed-tensors` (#6524) · 889da130
  Robert Shaw authored Jul 25, 2024
```
Co-authored-by: mgoin <michael@neuralmagic.com>
```
  889da130
23 Jul, 2024 2 commits
- [Misc] Add ignored layers for `fp8` quantization (#6657) · 0eb0757b
  Michael Goin authored Jul 23, 2024
  
  0eb0757b
- [Misc] Support FP8 kv cache scales from compressed-tensors (#6528) · 9e0b558a
  Michael Goin authored Jul 23, 2024
  
  9e0b558a
20 Jul, 2024 1 commit
- [ Misc ] `fbgemm` checkpoints (#6559) · 683e3cb9
  Robert Shaw authored Jul 20, 2024
  
  683e3cb9
19 Jul, 2024 1 commit
- [ Kernel ] Enable Dynamic Per Token `fp8` (#6547) · 4cc24f01
  Robert Shaw authored Jul 19, 2024
  
  4cc24f01
16 Jul, 2024 1 commit
- [Kernel][Attention] Separate `Attention.kv_scale` into `k_scale` and `v_scale` (#6081) · 978aed53
  Michael Goin authored Jul 16, 2024
  
  978aed53
14 Jul, 2024 1 commit
- [ Misc ] Apply MoE Refactor to Deepseekv2 To Support Fp8 (#6417) · fb6af8bc
  Robert Shaw authored Jul 13, 2024
  
  fb6af8bc
11 Jul, 2024 1 commit
- [ Misc ] Refactor Marlin Python Utilities (#6082) · b675069d
  Robert Shaw authored Jul 11, 2024
```
Co-authored-by: Robert Shaw <rshaw@neuralmagic.com>
```
  b675069d
07 Jul, 2024 1 commit
- [ Misc ] Support Fp8 via `llm-compressor` (#6110) · abfe705a
  Robert Shaw authored Jul 07, 2024
```
Co-authored-by: Robert Shaw <rshaw@neuralmagic>
```
  abfe705a
03 Jul, 2024 2 commits
- [Kernel] Expand FP8 support to Ampere GPUs using FP8 Marlin (#5975) · 47f0954a
  Michael Goin authored Jul 03, 2024
  
  47f0954a
- [hardware][misc] introduce platform abstraction (#6080) · 482045ee
  youkaichao authored Jul 02, 2024
  
  482045ee
02 Jul, 2024 1 commit

[ Misc ] Refactor MoE to isolate Fp8 From Mixtral (#5970) · 7c008c51

Robert Shaw authored Jul 02, 2024


Co-authored-by: Robert Shaw <rshaw@neuralmagic>
Co-authored-by: Michael Goin <michael@neuralmagic.com>

7c008c51

01 Jul, 2024 1 commit
- [misc][cuda] use nvml to avoid accidentally cuda initialization (#6007) · 614aa512
  youkaichao authored Jun 30, 2024
  
  614aa512
30 Jun, 2024 1 commit
- [ Misc ] Refactor w8a8 to use `process_weights_after_load` (Simplify Weight Loading) (#5940) · af9ad46f
  Robert Shaw authored Jun 30, 2024
```
Co-authored-by: Robert Shaw <rshaw@neuralmagic>
```
  af9ad46f
28 Jun, 2024 1 commit
- [ Bugfix ] Enabling Loading Models With Fused QKV/MLP on Disk with FP8 (#5921) · 2cd402e1
  Robert Shaw authored Jun 28, 2024
```
Co-authored-by: Robert Shaw <rshaw@neuralmagic>
```
  2cd402e1
20 Jun, 2024 1 commit
- [Bugfix] Fix the CUDA version check for FP8 support in the CUTLASS kernels (#5715) · 3f3b6b21
  Tyler Michael Smith authored Jun 20, 2024
  
  3f3b6b21
14 Jun, 2024 1 commit
- [Kernel] Fix CUTLASS 3.x custom broadcast load epilogue (#5516) · 703475f6
  Tyler Michael Smith authored Jun 14, 2024
  
  703475f6