[Refactor] refactor humming linear and moe backends to use explicit layer configs (#49610)

Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
Co-authored-by: OpenAI Codex <noreply@openai.com>
This commit is contained in:
Jinzhen Lin
2026-08-07 10:03:19 -07:00
committed by GitHub
co-authored by OpenAI Codex
parent e229fdbc87
commit fcde8e1460
37 changed files with 448 additions and 341 deletions
+6 -2
View File
@@ -15,9 +15,13 @@ _EXPORTS: dict[str, str] = {
"dtypes": "humming.dtypes",
"DataType": "humming.dtypes:DataType",
"GemmType": "humming.config:GemmType",
"LayerConfig": "humming.config:LayerConfig",
"WeightScaleType": "humming.config:WeightScaleType",
"HummingMethod": "humming.layer:HummingMethod",
"HummingLayerMeta": "humming.layer:HummingLayerMeta",
"humming_forward": "humming.forward:humming_forward",
"may_quant_input": "humming.forward:may_quant_input",
"prepare_layer_config": "humming.transform:prepare_layer_config",
"transform_humming_tensors": "humming.transform:transform_humming_tensors",
"get_heuristics_config": "humming.tune:get_heuristics_config",
"BaseInputSchema": "humming.schema:BaseInputSchema",
"BaseWeightSchema": "humming.schema:BaseWeightSchema",
"HummingInputSchema": "humming.schema:HummingInputSchema",