JunnanLi commited on
Commit
db2287f
1 Parent(s): afc6a0f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -1
README.md CHANGED
@@ -35,11 +35,25 @@ This allows the model to be used for tasks like:
35
  - visual question answering (VQA)
36
  - chat-like conversations by feeding the image and the previous conversation as prompt to the model
37
 
38
- ## Intended uses & limitations
39
 
40
  You can use the raw model for conditional text generation given an image and optional text. See the [model hub](https://huggingface.co/models?search=Salesforce/blip) to look for
41
  fine-tuned versions on a task that interests you.
42
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
43
  ### How to use
44
 
45
  For code examples, we refer to the [documentation](https://huggingface.co/docs/transformers/main/en/model_doc/blip-2#transformers.Blip2ForConditionalGeneration.forward.example).
 
35
  - visual question answering (VQA)
36
  - chat-like conversations by feeding the image and the previous conversation as prompt to the model
37
 
38
+ ## Direct Use and Downstream Use
39
 
40
  You can use the raw model for conditional text generation given an image and optional text. See the [model hub](https://huggingface.co/models?search=Salesforce/blip) to look for
41
  fine-tuned versions on a task that interests you.
42
 
43
+ ## Bias, Risks, Limitations, and Ethical Considerations
44
+
45
+ BLIP2-OPT uses off-the-shelf OPT as the language model. It inherits the same risks and limitations as mentioned in Meta's model card.
46
+
47
+ > Like other large language models for which the diversity (or lack thereof) of training
48
+ > data induces downstream impact on the quality of our model, OPT-175B has limitations in terms
49
+ > of bias and safety. OPT-175B can also have quality issues in terms of generation diversity and
50
+ > hallucination. In general, OPT-175B is not immune from the plethora of issues that plague modern
51
+ > large language models.
52
+ >
53
+ BLIP2 is fine-tuned on image-text datasets (e.g. [LAION](https://laion.ai/blog/laion-400-open-dataset/) ) collected from the internet. As a result the model itself is potentially vulnerable to generating equivalently inappropriate content or replicating inherent biases in the underlying data.
54
+
55
+ BLIP2 has not been tested in real world applications. It should not be directly deployed in any applications. Researchers should first carefully assess the safety and fairness of the model in relation to the specific context they’re being deployed within.
56
+
57
  ### How to use
58
 
59
  For code examples, we refer to the [documentation](https://huggingface.co/docs/transformers/main/en/model_doc/blip-2#transformers.Blip2ForConditionalGeneration.forward.example).