Hugging Face's logo

OPEA
/
Llama-3.3-70B-Instruct-int2-sym-inc

Safetensors
llama
2-bit
intel/auto-round
Model card Files Community
cicdatopea commited on
Commit
c35f89c
·
verified ·
1 Parent(s): 90984ae

revert the change

Browse files
Files changed (1) hide show
  1. README.md +1 -2
README.md CHANGED
@@ -120,9 +120,8 @@ lm-eval --model hf --model_args pretrained=OPEA/Llama-3.3-70B-Instruct-int2-sym-
120
  | arc_easy | 0.8624 | 0.8367 | | 0.8064 |
121
  | leaderboard_ifeval | 0.6661=(0.7110+0.6211)/2 | 0.61235=(0.6739+0.5508)/2 | | (0.5959+0.4603)/2 |
122
  | arc_challenge | 0.6109 | 0.5580 | | 0.5188 |
123
- | gsm8k(5shot) strict match? | 0.9083 | 0.8575 | | |
124
 
125
- gsm8k data is not correct, probably flexible-extract
126
 
127
  ## Generate the model
128
 
 
120
  | arc_easy | 0.8624 | 0.8367 | | 0.8064 |
121
  | leaderboard_ifeval | 0.6661=(0.7110+0.6211)/2 | 0.61235=(0.6739+0.5508)/2 | | (0.5959+0.4603)/2 |
122
  | arc_challenge | 0.6109 | 0.5580 | | 0.5188 |
123
+ | gsm8k(5shot) strict match | 0.9083 | 0.8575 | | |
124
 
 
125
 
126
  ## Generate the model
127