Update README.md
Browse files
README.md
CHANGED
|
@@ -4,6 +4,11 @@ pipeline_tag: visual-question-answering
|
|
| 4 |
---
|
| 5 |
# 😈 Imp
|
| 6 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
\[Technical report (coming soon)\] [[Demo](https://xmbot,net/imp/)\] [[Github](https://github.com/MILVLG/imp)\]
|
| 8 |
|
| 9 |
## Introduction
|
|
@@ -53,7 +58,7 @@ print(tokenizer.decode(output_ids[input_ids.shape[1]:], skip_special_tokens=True
|
|
| 53 |
## Model evaluation
|
| 54 |
We conduct evaluation on 9 commonly-used benchmarks, including 5 academic VQA benchmarks and 4 popular MLLM benchmarks, to compare our Imp model with LLaVA (7B) and existing MSLMs of similar model sizes.
|
| 55 |
|
| 56 |
-
| Models | Size | VQAv2 | GQA |VizWiz | SQA
|
| 57 |
|:--------:|:-----:|:----:|:----:|:-------------:|:--------:|:-----:|:----:|:-------:|:-------:|:-------:|
|
| 58 |
| [LLaVA-v1.5-lora](https://huggingface.co/liuhaotian/llava-v1.5-7b) | 7B |79.10 | **63.00** |47.80 | 68.40 |58.20| 86.40 | **1476.9** | 66.10 |30.2|
|
| 59 |
| [TinyGPT-V](https://huggingface.co/Tyrannosaurus/TinyGPT-V) | 3B | - | 33.60 | 24.80 | - | - | -| - | - |-|
|
|
|
|
| 4 |
---
|
| 5 |
# 😈 Imp
|
| 6 |
|
| 7 |
+
> A very small man can cast a very large shadow.
|
| 8 |
+
>
|
| 9 |
+
> ——*George R.R. Martin, A Clash of Kings*
|
| 10 |
+
|
| 11 |
+
|
| 12 |
\[Technical report (coming soon)\] [[Demo](https://xmbot,net/imp/)\] [[Github](https://github.com/MILVLG/imp)\]
|
| 13 |
|
| 14 |
## Introduction
|
|
|
|
| 58 |
## Model evaluation
|
| 59 |
We conduct evaluation on 9 commonly-used benchmarks, including 5 academic VQA benchmarks and 4 popular MLLM benchmarks, to compare our Imp model with LLaVA (7B) and existing MSLMs of similar model sizes.
|
| 60 |
|
| 61 |
+
| Models | Size | VQAv2 | GQA |VizWiz | SQA(IMG) | TextVQA | POPE | MME(P) | MMB |MM-Vet|
|
| 62 |
|:--------:|:-----:|:----:|:----:|:-------------:|:--------:|:-----:|:----:|:-------:|:-------:|:-------:|
|
| 63 |
| [LLaVA-v1.5-lora](https://huggingface.co/liuhaotian/llava-v1.5-7b) | 7B |79.10 | **63.00** |47.80 | 68.40 |58.20| 86.40 | **1476.9** | 66.10 |30.2|
|
| 64 |
| [TinyGPT-V](https://huggingface.co/Tyrannosaurus/TinyGPT-V) | 3B | - | 33.60 | 24.80 | - | - | -| - | - |-|
|