Refactor AWQ Marlin MoE onto modular WNA16 oracle (#42483)

Signed-off-by: Siddharth Bedekar <bedeksid@gmail.com>
Signed-off-by: Siddharth Bedekar <104613085+bedeks@users.noreply.github.com>
Co-authored-by: Robert Shaw <robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: OpenAI Codex <codex@openai.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
This commit is contained in:
Siddharth Bedekar
2026-05-18 08:02:43 -07:00
committed by GitHub
co-authored by Robert Shaw OpenAI Codex gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
parent f5d3dc7115
commit 78e7a7b9b0
6 changed files with 405 additions and 206 deletions
@@ -659,6 +659,8 @@ class AutoGPTQMoEMethod(FusedMoEMethodBase):
w2_g_idx,
w13_g_idx_sort_indices,
w2_g_idx_sort_indices,
_w13_qzeros,
_w2_qzeros,
w13_input_global_scale,
w2_input_global_scale,
w13_bias,
@@ -729,7 +731,6 @@ class AutoGPTQMoEMethod(FusedMoEMethodBase):
moe_quant_config=self.moe_quant_config,
moe_config=self.moe,
experts_cls=self.experts_cls,
layer=layer,
is_k_full=self.is_k_full,
w13_g_idx=layer.w13_g_idx,
w2_g_idx=layer.w2_g_idx,