/f/MachineLearning

Submissions
Comments

- Hot
- New
- Active
- Top
- Controversial
- Most commented

[D] Is validation set necessary for non-neural network models, too?

Submitted by osedao t3_11ayjxt on February 24, 2023 at 6:38 PM

15 comments

[D] What is the correct term for a non-GAN system where two or more networks compete as part of training?

Submitted by mosquitoLad t3_11au29r on February 24, 2023 at 3:29 PM

16 comments

[D] Best Way to Measure LLM Uncertainty?

Submitted by _atswi_ t3_11aq4qo on February 24, 2023 at 12:21 PM

8 comments

[D] A funny story from my interview

Submitted by nobody0014 t3_11albpf on February 24, 2023 at 7:16 AM

12 comments

[P] Minds - A JS library to build LLM powered backends and workflows (OpenAI & Cohere)

Submitted by gsvclass t3_11ak97p on February 24, 2023 at 6:10 AM

12 comments

[D] Can ML be useful in spectra analysis?

Submitted by NotSoChildishRubino t3_11addy8 on February 24, 2023 at 12:20 AM

7 comments

[D] To the ML researchers and practitioners here, do you worry about AI safety/alignment of the type Eliezer Yudkowsky describes?

Submitted by SchmidhuberDidIt t3_11ada91 on February 24, 2023 at 12:16 AM

176 comments

123

[P] What are the latest "out of the box solutions" for deploying the very large LLMs as API endpoints?

Submitted by johnhopiler t3_11a8tru on February 23, 2023 at 9:09 PM

11 comments

[D] Are there any good FID and KID metrics implementations existing that are compatible with pytorch?

Submitted by ats678 t3_11a4klg on February 23, 2023 at 6:16 PM

7 comments

[D] Model size vs task complexity

Submitted by Fine-Topic-6127 t3_119ydqv on February 23, 2023 at 1:59 PM

6 comments

[D] Tools for drawing/visualising Neural Networks that are pretty?

Submitted by CHvader t3_119w7c6 on February 23, 2023 at 12:11 PM

6 comments

[D] 14.5M-15M is the smallest number of parameters I could find for current pretrained language models. Are there any that are smaller?

Submitted by Seankala t3_119onf8 on February 23, 2023 at 4:41 AM

12 comments

[D] Python library to collect structured datasets across the internet

Submitted by dmart89 t3_119o54q on February 23, 2023 at 4:15 AM

17 comments

[N] U.S. Copyright Office decides that Kris Kashtanova's AI-involved graphic novel will remain copyright registered, but the copyright protection will be limited to the text and the whole work as a compilation

Submitted by Wiskkey t3_1198k5j on February 22, 2023 at 7:19 PM

6 comments

[P] MIT Introduction to Data-Centric AI

Submitted by anishathalye t3_1194wm0 on February 22, 2023 at 5:00 PM

8 comments

375

[D] Faster Flan-T5 inference

Submitted by _learn_faster_ t3_1194vcc on February 22, 2023 at 4:59 PM

8 comments

[R] Provable Copyright Protection for Generative Models

Submitted by vyasnikhil96 t3_1190lw8 on February 22, 2023 at 2:22 PM

29 comments

[D] Training a UNet-like architecture for semantic segmentation with 200 outcome classes.

Submitted by Scared_Employer6992 t3_11dd59q on February 27, 2023 at 2:28 PM

10 comments

[D] "Deep learning is the only thing that currently works at scale"

Submitted by GraciousReformer t3_118pof6 on February 22, 2023 at 5:12 AM

117 comments

122

[Discussion] Exploring the Black Box Theory and Its Implications for AI, God, and Ethics

Submitted by Disastrous_Nose_1299 t3_118ovfo on February 22, 2023 at 4:29 AM

57 comments

[D] Stable Diffusion, Class images question

Submitted by vurt72 t3_118n22u on February 22, 2023 at 2:56 AM

7 comments

Unit Normalization instead of Cross-Entropy Loss [Discussion]

Submitted by thomasahle t3_118gie9 on February 21, 2023 at 10:04 PM

7 comments

[D] Bottleneck Layers: What's your intuition?

Submitted by _Arsenie_Boca_ t3_118cypl on February 21, 2023 at 7:22 PM

16 comments

[P] The First Depthwise-separable Convolution Animation

Submitted by Animated-AI t3_118c8pp on February 21, 2023 at 6:23 PM

23 comments

338

[R] ChatGPT for Robotics: Design Principles and Model Abilities

Submitted by CheapBreakfast9 t3_11853g5 on February 21, 2023 at 2:46 PM

8 comments