Commit
•
efde4fc
1
Parent(s):
858da5f
Update README.md
Browse files
README.md
CHANGED
@@ -11,10 +11,12 @@ pipeline_tag: image-to-text
|
|
11 |
base_model: Salesforce/blip2-opt-2.7b
|
12 |
---
|
13 |
# VLRM
|
14 |
-
This repository contains the weights of BLIP-2 OPT-2.7B model fine-tuned by reinforcement learning method introduced in the paper VLRM: Vision-Language Models
|
15 |
Reward Models for Image Captioning.
|
16 |
|
17 |
The RL-tuned model is able to generate longer and more comprehensive descriptions with zero computational overhead compared to the original model.
|
|
|
|
|
18 |
# Running the model
|
19 |
## Option 1
|
20 |
<details>
|
|
|
11 |
base_model: Salesforce/blip2-opt-2.7b
|
12 |
---
|
13 |
# VLRM
|
14 |
+
This repository contains the weights of BLIP-2 OPT-2.7B model fine-tuned by reinforcement learning method introduced in the paper VLRM: Vision-Language Models Act as
|
15 |
Reward Models for Image Captioning.
|
16 |
|
17 |
The RL-tuned model is able to generate longer and more comprehensive descriptions with zero computational overhead compared to the original model.
|
18 |
+
# CLIP Recall
|
19 |
+
CLIP Recall calculation scripts are provided in `validate` folder together with `README.md` and `captions.txt`.
|
20 |
# Running the model
|
21 |
## Option 1
|
22 |
<details>
|