Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Corrective Instruction Generation on HumanML3D (test)
Loading...
100
BLEU
Ground-Truth
-1.92
24.54
51
77.46
Dec 6, 2024
BLEU
ROUGE
METEOR
CLIPScore
Updated 4d ago
Evaluation Results
Method
Method
Links
BLEU
ROUGE
METEOR
CLIPScore
Ground-Truth
2024.12
100
100
100
1
CigTime
Mode=Full-parameter tu...
2024.12
24
35
52
0.82
Mistral-7B
Mode=In-context learning
2024.12
16
30
46
0.8
Llama-3-8B
Mode=In-context learning
2024.12
15
29
45
0.77
Llama-3-8B-LoRA
Mode=LoRA fine-tuning
2024.12
10
19
36
0.77
Mistral-7B-LoRA
Mode=LoRA fine-tuning
2024.12
8
19
27
0.79
MotionGPT
Input Template=Section...
2024.12
2
10
11
0.76
MotionGPT-M2T
Description=Motion cap...
2024.12
2
13
12
0.76
Feedback
Search any
task
Search any
task