Skip to content
The methodOpen data
Back to the collection

BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models

Junnan Li, Dongxu Li, Silvio Savarese, Steven Hoi

2023 · arXiv

Language modelsDatabases
Read the paper
HN score
0
Mentions
3
Readers
3
Months
3
What’s in a score?

From the abstract

The cost of vision-and-language pre-training has become increasingly prohibitive due to end-to-end training of large-scale models. This paper proposes BLIP-2, a generic and efficient pre-training strategy that bootstraps vision-language pre-training from off-the-shelf frozen pre-trained image encoders and frozen large language models. BLIP-2 bridges the modality gap with a lightweight Querying Transformer, which is pre-trained in two stages. The first stage bootstraps vision-language representation learning from a frozen image encoder. The second stage bootstraps vision-to-language generative learning from a frozen language model. BLIP-2 achieves state-of-the-art performance on various vision-language tasks, despite having significantly fewer trainable parameters than existing methods. For example, our model outperforms Flamingo80B by 8.7% on zero-shot VQAv2 with 54x fewer trainable parameters. We also demonstrate the model's emerging capabilities of zero-shot image-to-text generation that can follow natural language instructions.

Metadata: arXiv. Abstracts retain their original authors’ rights.

In the conversation

Original HN comments, with labels from simple text rules. A label describes the whole comment and may miss context or irony. Read the discussion to make up your own mind. About the labels

3 selected excerpts · 3 total mentions

HarHarVeryFunny
Neutral mention · rule
It has more in common with GPT-3 than GPT-4 in terms of size, but in reality it's based on Vicuna/Llama which is 10x smaller than either, so as far as the LLM part of it goes its not mini-anything - it's just straight-up Vicuna 13B. The model as a whole is just BLIP-2 with a larger linear layer, and using Vicuna as the LLM. If you look at their code it's literally using the entire BLIP-2 encoder (Salesforce code). https://arxiv.org/pdf/2301.12597.pdf
Original comment MiniGPT-4
f38zf5vdt
Neutral mention · rule
You are humanizing token prediction. The multimodal models for text-vision were all established using a scaffold of architectures that unified text-token and vision-token similarity e.g. BLIP2. [1] It's possible that a model using unified representations might be able to establish that the set of visual tokens you are searching for corresponds to some set of text tokens, but only if the pretrained weights for the vision encoder are able to extract the features corresponding to the object to which you are describing to the vision model. And the pretrained vision encoder will have at some point been trained to minimize text-visual token cosine similarity on some training set, so it really depends on what exactly that training set had in it. [1] https://arxiv.org/pdf/2301.12597.pdf
Original comment Zero-1-to-3: Zero-shot One Image to 3D Object

Keep following the thread.

More in Language models

More from the same topics.

2017arXiv

The Case for Learned Index Structures

Tim Kraska, Alex Beutel, Ed H. Chi +2

Indexes are models: a B-Tree-Index can be seen as a model to map a key to the position of a record within a sorted array, a Hash-Index as a model to map a key to a position of a record within an unsorted array, and a BitMap-Index as a model to indicate if a data record exists or not.

Databases
19 mentions14 months
5HN score