Multimodal Autoregressive Pre-training of Large Vision Encoders
Paper
•
2411.14402
•
Published
•
32
Fynn Kröger
fynnkroeger
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
about 6 hours ago
Multimodal Autoregressive Pre-training of Large Vision Encoders
upvoted
a
paper
11 days ago
Edify Image: High-Quality Image Generation with Pixel Space Laplacian
Diffusion Models
upvoted
a
paper
27 days ago
Breaking the Memory Barrier: Near Infinite Batch Size Scaling for
Contrastive Loss
Organizations
None yet
fynnkroeger's activity
upvoted
a
paper
about 6 hours ago
upvoted
a
paper
11 days ago
upvoted
a
paper
27 days ago
upvoted
a
paper
about 1 month ago
liked
a
model
about 1 month ago
liked
a
Space
about 1 month ago
liked
a
Space
about 2 months ago
upvoted
2
papers
about 2 months ago
upvoted
a
paper
2 months ago
Reacted to
davidberenstein1957's
post with 🤗
2 months ago
Post
1342
Distilabel and synthetic data community interviews - the outcomes
We've been doing some interview with community members to understand the needs surrounding synthetic data. Many thanks to the participants. Note that, given they interviewees were sourced from our community, so the results will likely represent that.
Things distilabel does well
- security and reliability by caching generations and having serializable pipelines.
- scaling up generation by parallelising inference and Anyscale Ray
- solid implementations of state of the art research papers
Things to improve
- communication about the fact we support structured generation
- customization of existing prompt implementations are difficult
- creation of new tasks prove difficult
- arguments and parameters for tasks aren't available at first glance
- the learning curve can be steep
- more tutorials that represent real-life usage
Things to note
- create small scale and large scale dataset to Millions of records
- people use synthetic data to move away from frontier model providers
- people mostly use 7B or 70B models for generating
Participate here: https://github.com/argilla-io/distilabel/issues
We've been doing some interview with community members to understand the needs surrounding synthetic data. Many thanks to the participants. Note that, given they interviewees were sourced from our community, so the results will likely represent that.
Things distilabel does well
- security and reliability by caching generations and having serializable pipelines.
- scaling up generation by parallelising inference and Anyscale Ray
- solid implementations of state of the art research papers
Things to improve
- communication about the fact we support structured generation
- customization of existing prompt implementations are difficult
- creation of new tasks prove difficult
- arguments and parameters for tasks aren't available at first glance
- the learning curve can be steep
- more tutorials that represent real-life usage
Things to note
- create small scale and large scale dataset to Millions of records
- people use synthetic data to move away from frontier model providers
- people mostly use 7B or 70B models for generating
Participate here: https://github.com/argilla-io/distilabel/issues
Reacted to
TuringsSolutions's
post with 😔
3 months ago
Post
1315
I can solve the Traveling Salesman Problem using the same methods the scientists used to solve it with 1 qubit, except I do not need quantum computers to do it. I am kind of tired of screaming this from the rooftops at this point. I can create an imaginary probability space, then I can put a bunch of imaginary agents in the imaginary box, and solve real problems in seconds. Problems that would take minutes, hours, or years to solve via other algorithms. Here is a demo of me solving the Traveling Salesman problem using 50 agents to probabilistically sample at once: https://colab.research.google.com/drive/1XplG72nQDO_-2h4DUllERLp0Dr2pI2J2?usp=sharing
- 28 replies
upvoted
8
papers
3 months ago
OLMoE: Open Mixture-of-Experts Language Models
Paper
•
2409.02060
•
Published
•
77
VisionTS: Visual Masked Autoencoders Are Free-Lunch Zero-Shot Time Series Forecasters
Paper
•
2408.17253
•
Published
•
35
Law of Vision Representation in MLLMs
Paper
•
2408.16357
•
Published
•
92
Scalable Autoregressive Image Generation with Mamba
Paper
•
2408.12245
•
Published
•
25
Towards Conversational Diagnostic AI
Paper
•
2401.05654
•
Published
•
16
MegaFusion: Extend Diffusion Models towards Higher-resolution Image Generation without Further Tuning
Paper
•
2408.11001
•
Published
•
11
Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model
Paper
•
2408.11039
•
Published
•
56
JPEG-LM: LLMs as Image Generators with Canonical Codec Representations
Paper
•
2408.08459
•
Published
•
44