Hugging Face's logo

PhoenixZ
/
LLaVANext-OmniAlign-32B-DPO

Safetensors
llava_qwen
Model card Files Community
1
PhoenixZ commited on
Commit
61ce6a8
·
verified ·
1 Parent(s): 85dfb96

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -10,7 +10,7 @@ SFT Dataset: [OmniAlign-V](https://huggingface.co/datasets/PhoenixZ/OmniAlign-V)
10
 
11
  DPO Dataset: [OmniAlign-V-DPO](https://huggingface.co/datasets/PhoenixZ/OmniAlign-V-DPO),
12
 
13
- MM-AlignBench: [MM-AlignBench](https://github.com/open-compass/VLMEvalKit)
14
 
15
  Checkpoints: [LLaVANext-OA-7B](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-7B), [LLaVANext-OA-32B](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-32B), [LLaVANext-OA-32B-DPO](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-32B-DPO)
16
 
@@ -25,9 +25,9 @@ By integrating OmniAlign-V-DPO datasets in DPO stage, we can further improve the
25
 
26
  | Model | Win Rate | Reward | Better+ | Better | Tie | Worse | Worse+ |
27
  |-------------------------------|------------------------------|---------------------------|------------|-----|----|-----|-----|
28
- | Claude3.5V-Sonnet | 84.9 | +51.4 | 70 | 144 | 12 | 31 | 4 |
29
  | GPT-4o | 81.3 | +49.0 | 81 | 124 | 12 | 31 | 4 |
30
- | GPT-4V | 82.5 | +46.0 | 57 | 157 | 12 | 31 | 1 |
31
  | GeminiFlash1.5-002 | 77.0 | +39.1 | 56 | 138 | 14 | 35 | 9 |
32
  | LLaVANext-OA-32B-DPO | 74.2 | +36.9 | 49 | 138 | 20 | 40 | 5 |
33
  | Qwen2VL-72B | 61.5 | +21.6 | 43 | 112 | 15 | 75 | 7 |
 
10
 
11
  DPO Dataset: [OmniAlign-V-DPO](https://huggingface.co/datasets/PhoenixZ/OmniAlign-V-DPO),
12
 
13
+ MM-AlignBench: [VLMEvalkit](https://github.com/open-compass/VLMEvalKit), [Huggingface](https://huggingface.co/datasets/PhoenixZ/MM-AlignBench)
14
 
15
  Checkpoints: [LLaVANext-OA-7B](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-7B), [LLaVANext-OA-32B](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-32B), [LLaVANext-OA-32B-DPO](https://huggingface.co/PhoenixZ/LLaVANext-OmniAlign-32B-DPO)
16
 
 
25
 
26
  | Model | Win Rate | Reward | Better+ | Better | Tie | Worse | Worse+ |
27
  |-------------------------------|------------------------------|---------------------------|------------|-----|----|-----|-----|
28
+ | Claude3.5V-Sonnet | 84.9 | +51.4 | 70 | 144 | 13 | 25 | 0 |
29
  | GPT-4o | 81.3 | +49.0 | 81 | 124 | 12 | 31 | 4 |
30
+ | GPT-4V | 82.5 | +46.0 | 57 | 151 | 12 | 31 | 1 |
31
  | GeminiFlash1.5-002 | 77.0 | +39.1 | 56 | 138 | 14 | 35 | 9 |
32
  | LLaVANext-OA-32B-DPO | 74.2 | +36.9 | 49 | 138 | 20 | 40 | 5 |
33
  | Qwen2VL-72B | 61.5 | +21.6 | 43 | 112 | 15 | 75 | 7 |