Skip to content

PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度

Merge branch 'main' of https://github.com/intel/auto-round into main

Zhang, Weiwei10557ed277b9d
diff --cc examples/language-modeling/main.py
index 7835995fa,4fa9bace4..7892922fc
--- a/examples/language-modeling/main.py
+++ b/examples/language-modeling/main.py
@@@ -101,8 -105,11 +105,12 @@@ if __name__ == '__main__'
parser.add_argument("--output_dir", default="./tmp_autoround", type=str,
help="Where to store the final model.")
+
+ parser.add_argument("--eval_legacy", action='store_true',
+ help="Whether to evaluate with a old lm_eval version(e.g. 0.3.0).")
+
+
args = parser.parse_args()
set_seed(args.seed)