初始化项目,由ModelHub XC社区提供模型

Model: neuralmagic/starcoder2-3b-quantized.w8a8
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-28 11:38:15 +08:00
commit 0a10c725f6
12 changed files with 148047 additions and 0 deletions

8
recipe.yaml Normal file
View File

@@ -0,0 +1,8 @@
quant_stage:
quant_modifiers:
GPTQModifier:
sequential_update: false
dampening_frac: 0.01
ignore: [lm_head]
scheme: W8A8
targets: Linear