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

remove gpt ppl eval from lm-0.4.2 (#158)

wenhuach2177d6a886d0c3
diff --git a/examples/language-modeling/main.py b/examples/language-modeling/main.py
index 3ef1fb452..d5171fe55 100644
--- a/examples/language-modeling/main.py
+++ b/examples/language-modeling/main.py
@@ -94,7 +94,7 @@ if __name__ == '__main__':
parser.add_argument("--tasks",
default="lambada_openai,hellaswag,winogrande,piqa,mmlu,wikitext,truthfulqa_mc1," \
- "truthfulqa_mc2,openbookqa,boolq,rte,arc_easy,arc_challenge,wikitext2,ptb-new,c4-new",
+ "truthfulqa_mc2,openbookqa,boolq,rte,arc_easy,arc_challenge",
help="lm-eval tasks for lm_eval version 0.4")
parser.add_argument("--output_dir", default="./tmp_autoround", type=str,
@@ -221,13 +221,6 @@ if __name__ == '__main__':
if args.model_dtype == "bfloat16" or args.model_dtype == "bfp16":
model = model.to(torch.bfloat16)
- # if "llama" in model_name:
- # from transformers import LlamaTokenizer
- #
- # tokenizer = LlamaTokenizer.from_pretrained(model_name)
- # if tokenizer.pad_token is None:
- # tokenizer.add_special_tokens({'pad_token': '[PAD]'})
- # else:
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=not args.disable_trust_remote_code)
if hasattr(tokenizer, "model_max_length"):