Crossing Linguistic Horizons

Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models

Few-Shot Language Modeling Leaderboard

Models MLQA-MLM VSEC
EM CER WER CED WED PLX EM CER WER CED WED PLX
URA-LLaMa 70B 0.01 ± 0.00 0.54 ± 0.00 0.66 ± 0.00 669.74 ± 10.38 153.04 ± 2.33 1.32 ± 0.05 0.33 ± 0.00 0.11 ± 0.00 0.13 ± 0.00 15.09 ± 0.42 4.05 ± 0.11 1.13 ± 0.00
URA-LLaMa 13B 0.01 ± 0.00 0.45 ± 0.01 0.61 ± 0.01 559.64 ± 11.23 136.97 ± 2.68 1.49 ± 0.10 0.35 ± 0.00 0.02 ± 0.00 0.04 ± 0.00 2.81 ± 0.12 1.18 ± 0.03 1.15 ± 0.00
URA-LLaMa 7B 0.01 ± 0.00 0.40 ± 0.01 0.55 ± 0.01 498.36 ± 11.01 118.11 ± 2.58 1.24 ± 0.01 0.22 ± 0.00 0.32 ± 0.01 0.33 ± 0.01 41.89 ± 1.54 10.10 ± 0.34 1.07 ± 0.00
LLaMa-2 13B 0.01 ± 0.00 0.74 ± 0.00 0.87 ± 0.00 760.98 ± 11.91 186.90 ± 2.85 1.24 ± 0.03 0.16 ± 0.00 0.03 ± 0.00 0.05 ± 0.00 3.38 ± 0.16 1.51 ± 0.04 1.01 ± 0.00
LLaMa-2 7B 0.00 ± 0.00 0.81 ± 0.00 0.98 ± 0.00 769.36 ± 10.51 198.53 ± 2.57 1.74 ± 0.19 0.12 ± 0.00 0.36 ± 0.01 0.39 ± 0.01 47.50 ± 0.86 11.80 ± 0.19 1.06 ± 0.00
Vietcuna 7B 0.00 ± 0.00 1.04 ± 0.00 1.06 ± 0.00 935.65 ± 12.47 204.98 ± 2.79 1.40 ± 0.00 0.00 ± 0.00 8.00 ± 0.07 8.01 ± 0.07 1063.93 ± 7.64 241.74 ± 1.74 1.46 ± 0.00
MixSUra 8x7B 0.00 ± - 0.55 ± - 0.63 ± - 526.79 ± - 131.02 ± - 1.00 ± - 0.08 ± - 0.19 ± - 0.28 ± - 25.13 ± - 8.58 ± - 1.00 ± -
GPT-3.5 0.04 ± 0.00 0.28 ± 0.01 0.44 ± 0.01 387.37 ± 10.86 92.78 ± 2.46 - 0.66 ± 0.00 0.01 ± 0.00 0.02 ± 0.00 1.63 ± 0.08 0.61 ± 0.02 -
GPT-4 0.08 ± 0.00 0.23 ± 0.01 0.40 ± 0.01 336.53 ± 10.18 83.55 ± 2.34 - 0.75 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 0.89 ± 0.04 0.37 ± 0.01 -