Spaces:
Runtime error
Runtime error
nitzanguetta
commited on
Commit
•
946cdc1
1
Parent(s):
b9a6953
results updated
Browse files
WHOOPS_Explanation_of_Violation_Leaderboard.tsv
CHANGED
|
@@ -1,7 +1,16 @@
|
|
| 1 |
Model Human Metric Auto Metric Identify (Binary Accuracy)
|
| 2 |
Humans 95 92
|
| 3 |
-
Ground-truth Caption
|
|
|
|
|
|
|
|
|
|
| 4 |
BLIP2 FlanT5-XXL (Fine-tuned) 27 24 73
|
| 5 |
BLIP2 FlanT5-XL (Fine-tuned) 15 18 60
|
| 6 |
-
Predicted Caption
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
BLIP2 FlanT5-XXL (Zero-shot) 0 0 50
|
|
|
|
| 1 |
Model Human Metric Auto Metric Identify (Binary Accuracy)
|
| 2 |
Humans 95 92
|
| 3 |
+
Ground-truth Caption → Llama-2-7b (Oracle) 71
|
| 4 |
+
Ground-truth Caption → Llama-2-13b (Oracle) 70
|
| 5 |
+
Ground-truth Caption → GPT4 (Oracle) 69
|
| 6 |
+
Ground-truth Caption → GPT3 (Oracle) 68 62 74
|
| 7 |
BLIP2 FlanT5-XXL (Fine-tuned) 27 24 73
|
| 8 |
BLIP2 FlanT5-XL (Fine-tuned) 15 18 60
|
| 9 |
+
Predicted Caption → Llama-2-7b 36
|
| 10 |
+
Predicted Caption → Llama-2-13b 36
|
| 11 |
+
Predicted Caption → GPT4 36
|
| 12 |
+
Predicted Caption → GPT3 33 42 59
|
| 13 |
+
InstructBLIP 31
|
| 14 |
+
LLaVA 31
|
| 15 |
+
mPLUG-Owl 24
|
| 16 |
BLIP2 FlanT5-XXL (Zero-shot) 0 0 50
|