- 17 Sep, 2024 1 commit
-
-
chenqianfzh authored
-
- 13 Sep, 2024 1 commit
-
-
youkaichao authored
-
- 11 Sep, 2024 1 commit
-
-
Li, Jiang authored
-
- 29 Aug, 2024 1 commit
-
-
chenqianfzh authored
-
- 27 Aug, 2024 1 commit
-
-
Dipika Sikka authored
Co-authored-by:ElizaWszola <eliza@neuralmagic.com>
-
- 21 Aug, 2024 1 commit
-
-
youkaichao authored
-
- 16 Aug, 2024 3 commits
-
-
Mor Zusman authored
-
jon-chuang authored
-
Michael Goin authored
-
- 14 Aug, 2024 1 commit
-
-
Kyle Sayers authored
-
- 13 Aug, 2024 1 commit
-
-
Kyle Sayers authored
-
- 07 Aug, 2024 2 commits
-
-
Michael Goin authored
-
Dipika Sikka authored
[Misc] Refactor linear layer weight loading; introduce `BasevLLMParameter` and `weight_loader_v2` (#5874)
-
- 05 Aug, 2024 1 commit
-
-
Isotr0py authored
Co-authored-by:Michael Goin <michael@neuralmagic.com>
-
- 01 Aug, 2024 1 commit
-
-
Michael Goin authored
-
- 30 Jul, 2024 1 commit
-
-
Tyler Michael Smith authored
-
- 25 Jul, 2024 1 commit
-
-
Michael Goin authored
-
- 23 Jul, 2024 3 commits
-
-
dongmao zhang authored
Co-authored-by: -
Michael Goin authored
-
Michael Goin authored
-
- 21 Jul, 2024 1 commit
-
-
Alexander Matveev authored
-
- 16 Jul, 2024 1 commit
-
-
Michael Goin authored
-
- 11 Jul, 2024 1 commit
-
-
Robert Shaw authored
Co-authored-by:Robert Shaw <rshaw@neuralmagic.com>
-
- 07 Jul, 2024 1 commit
-
-
Robert Shaw authored
Co-authored-by:Robert Shaw <rshaw@neuralmagic>
-
- 03 Jul, 2024 3 commits
-
-
Robert Shaw authored
-
Michael Goin authored
-
youkaichao authored
-
- 02 Jul, 2024 1 commit
-
-
Qubitium-ModelCloud authored
Co-authored-by:
ZX <zx@lbx.dev>
-
- 01 Jul, 2024 1 commit
-
-
youkaichao authored
-
- 30 Jun, 2024 1 commit
-
-
Robert Shaw authored
Co-authored-by:
-
- 25 Jun, 2024 1 commit
-
-
Dipika Sikka authored
-
- 19 Jun, 2024 1 commit
-
-
Dipika Sikka authored
[Misc] Add per channel support for static activation quantization; update w8a8 schemes to share base classes (#5650)
-
- 18 Jun, 2024 1 commit
-
-
Dipika Sikka authored
[Misc] Add channel-wise quantization support for w8a8 dynamic per token activation quantization (#5542)
-
- 17 Jun, 2024 1 commit
-
-
Dipika Sikka authored
-
- 16 Jun, 2024 1 commit
-
-
Michael Goin authored
-
- 15 Jun, 2024 1 commit
-
-
Cyrus Leung authored
-
- 13 Jun, 2024 2 commits
-
-
Michael Goin authored
-
Dipika Sikka authored
Co-authored-by:Robert Shaw <114415538+robertgshaw2-neuralmagic@users.noreply.github.com>
-
- 12 Jun, 2024 2 commits
-
-
Cody Yu authored
Inspired by #5146, this PR improves FP8 quantize kernel by vectorizing data transfer to better utilize memory bandwidth. Microbenchmark shows that this improved kernel can achieve 1.0x-1.5x speedup (especially when hidden size is large). In details, we applied 3 optimizations: - Use inverted scale so that most divisions are changed to multiplications. - Unroll the loop by 4 times to improve ILP. - Use vectorized 4 to transfer data between HBM and SRAM.
-
Simon Mo authored
Revert "[CI/Build] Add `is_quant_method_supported` to control quantization test configurations" (#5463)
-
