Questions tagged [generative-models]
For questions about models designed for generating new data (or generating samples from a probability distribution).
187
questions
0
votes
1
answer
8
views
What methods exist for synthetic series data with entities?
By "entitiy" I mean the generative source of a single series. For example, in a dataset tracking sales of various items across multiple stores, each store would be an entity. Each one can ...
- 13
0
votes
0
answers
18
views
Understanding the embeddings model (dunzhang/stella_en_400M_v5) by Alibaba. The details about the retrieve task and the s2s task
The model I am talking about is hosted here:
From the documentation:
We simplify usage of prompts, providing two prompts for most general tasks, one is for s2p, another one is for s2s.Prompt of s2p ...
- 883
0
votes
0
answers
12
views
Why is the FID score of my StyleGAN model increasing as training progresses
I am training a StyleGAN model and I output the FID score every 10k iterations, here are the scores for introducing and stabilising phase of layer 3 (16x16) images.
...
- 1
0
votes
0
answers
8
views
why autoregressive models can't do representation learning(unsupervised learning)?
I was going through the lecture-4 of the deep generative modelling course from Cornell. In the lecture slides it is mentions that autoregressive models cannot do representation learning(min: 5.23). ...
0
votes
0
answers
9
views
CGAN Training Issues: Discriminator Accuracy at 100% and Generator Loss at 0
I am trying to train a Conditional Generative Adversarial Network (CGAN) to generate synthetic leaf images. However, during training, my discriminator's accuracy quickly reaches 100%, and the ...
0
votes
0
answers
11
views
Accurate score function estimation using score-based diffusion models
My question is mainly related to the seminal paper by Song et al.: "Score-Based Generative Modeling through Stochastic Differential Equations". I would like to leverage their framework in ...
- 51
0
votes
0
answers
8
views
Diffusion Model consistency term derivation question
The consistency term of the diffusion model is written as:
$$\mathop{\mathbb{E_{q_\phi(x_{1:T}|x_0)}}} \left[\log\prod_{t=2}^T \frac{p(x_{t-1} | x_t)}{q_\phi(x_{t-1}|x_t, x_0)}\right]$$
$$= \sum_{t=2}^...
-1
votes
1
answer
9
views
WGAN generating images from the training data
Is it possible for gan to remember somehow training data distribution?
Or maybe somеthing leaks out when I calculate gradients?
...
- 43
0
votes
0
answers
31
views
Why is the generator not trained directly in GAN
When we build a GAN network we usually do the following:
Build a discriminator and compile it
Build a generator
Build a combined model generator+discriminator and compile it
Now for training we do ...
- 133
0
votes
0
answers
13
views
does DALL-E api use Microsoft account info while generating responses?
I have read some work regarding occupational gender bias in AI image generation and it seemed. According to my research, tools like DALL-E generated more images of men when trying out for images with ...
- 191
0
votes
0
answers
15
views
Why does the TensorFlow docs use a different GAN generator loss?
As per the original paper that introduced GANs, the generator loss is given as:
$$
L_{G} = L _{BCE}(\mathbf{\vec 0}, \mathbf{D}(\mathbf{G}(\mathbf{\vec z}))) = \log(1 - \mathbf{D}(\mathbf{G}(\mathbf{\...
2
votes
1
answer
161
views
Diffusion Models: Conditioning on Time vs. Noise Level
I am new to SE-Data Science, therefore I hope this is the right place to ask this rather theoretical question.
In diffusion models we usually have a time variable which determines the noise schedule (...
- 133
2
votes
0
answers
46
views
Training VAE in Latent Diffusion Models
When working with Latent Diffusion Models (LDMs), is it common practice to only train the U-Net component while leaving the VAE untrained? Additionally, does this approach apply when fine-tuning an ...
- 33
0
votes
0
answers
14
views
cGAN: Discriminator loss going to zero while Generator's going always up but the result is very good
I have a Conditional Generative Adversarial Network for Quantum State Tomography. The metrics I am monitoring during the training process are the losses and the Fidelity (the degree of similarity ...
- 33
1
vote
0
answers
13
views
About GANs specialization
I'm working in a binary classification problem. In general terms, I'm focused on using neural networks to classify datasets that are not huge and in order to address that problem I'm comparing ...
- 51
0
votes
0
answers
18
views
Generation time in stable cascade vs stable diffusion
According to the docs, Stable Cascade is meant to be faster than SD due to its architecture. However, when trying out using Python on my local machine, the results seem to be opposite. I am using a ...
- 191
1
vote
1
answer
31
views
How to ensure that the output of my generative model is uniform?
Generative models transform noise to data. However, is there a way to ensure the output of my generative model follows a specific distribution, say uniform?
More concretely, applying a scalar function ...
- 21
1
vote
0
answers
20
views
How to derive at the expectation equation given in the paper "Video Diffusion Models"?
In the paper Video Diffusion Models, Section 3.1 mentions the following equation:
$$ E_q[x^b|,z_t,x^a] = E_q[x^b|z_t] + (\frac{{\sigma}_t^2}{{\alpha}_t})\nabla_{z_t^b}\log q(x^a|z_t)$$, where $x^a, x^...
- 21
0
votes
0
answers
20
views
training data of stable cascade vs stable diffusion
So according to this: https://waxy.org/2022/08/exploring-12-million-of-the-images-used-to-train-stable-diffusions-image-generator/
Stable Diffusion was trained on data from CommonCrawler. I believe ...
- 191
0
votes
0
answers
22
views
stabilityai/stable-cascade vs runwayml/stable-diffusion-v1-5
What are the major differences in these text-to-image AI models:
stabilityai/stable-cascade
runwayml/stable-diffusion-v1-5
in terms of architecture and performance?
- 191
1
vote
0
answers
18
views
Best way to achieve High throughput with ChatGPT (and LLMs in general) APIs
I was wondering what the best way is to perform API calls to ChatGPT, and in general, with LLMs API providers, to achieve the highest throughput in a batch inference context.
Suppose you have 1000 ...
- 53
0
votes
0
answers
24
views
Image Generation Models
I am looking for a list of different image generation models and how I can test them? For example:
DALL-E (accessible via ChatGPT4)
Stable Diffusion (open source)
CLIP? - but idk how to access/test it....
- 191
0
votes
0
answers
78
views
How can I apply a technique from "ZHOU, Pei, et al. Self-Discover: Large Language Models Self-Compose Reasoning Structures. 2024" to my toy project?
I want to practice a technique from "ZHOU, Pei, et al. Self-Discover: Large Language Models Self-Compose Reasoning Structures. arXiv preprint arXiv:2402.03620, 2024." The article call it &...
0
votes
1
answer
52
views
Prompt Ops Alternatives
What are the main alternatives for prompt ops nowadays? By prompt ops, I mean a comprehensive solution for tracking prompt engineering experiments and also registering prompts in different stages, ...
- 53
0
votes
0
answers
27
views
Fine-tuned Stable Diffusion not converging with custom encoder
I'm currently fine-tuning a stable diffusion model for the task of dataset augmentation. I am training the model on 80k images from hte CelebA-attributes dataset, replacing the text encoder with ...
4
votes
1
answer
111
views
LLMs for text generation
We know that AI is rapidly growing. do we have any large language models (LLMs) to process images, pdf documents directly (fine-tune approach) for text generation tasks?
- 43
1
vote
0
answers
133
views
Generating quality synthetic tabular data - is it possible when one's dataset is extremely small?
I've got a dataset consisting of only 17 samples and 6 continuous features (all values in the dataset contain decimals, although 2 features exhibit categorical-ish behaviour). I'm looking at the ...
- 11
0
votes
0
answers
35
views
Details about Pre-existing knowledge database in RAG for LLMs
In RAG, As part of retriever model- we are retrieving the relevant information from external knowledge source (i.e. vector database) and this database is always updating with new updates.
In ...
- 43
1
vote
1
answer
24
views
Extraction of name from phonetic transcription
I have a use case where I want to extract the name from the phonetic transcription.
For example if the phonetic transcription is - “m a j n e j m ɪ z s ʌ m i ɹ z o w ʃ i”, the output should be the ...
- 111
0
votes
0
answers
9
views
Strategies for Encoding Large Datasets in Symbolic Music Generation for BERT-type Model
I am creating a BERT-type model for symbolic music generation. An observation of my database is a musical piece. Actually, is a "viewpoint" of the piece: ...
- 101
0
votes
0
answers
102
views
How to improve performance of a Retrieval Augmented Generative (RAG) model?
I had implemented a Retrieval Augmented generation (RAG) model on the Healthcare CSV file. The model has to give answers to natural language queries based on the data provided. After implementing the ...
0
votes
0
answers
13
views
Use OpenNN for score matching
I‘m new to this topic and need some guidance. I would like to implement score matching (like sliced-score matching: https://arxiv.org/abs/1905.07088) in C++. OpenNN (https://www.opennn.net/) seems to ...
- 101
0
votes
1
answer
165
views
How the retriever model (Query encoder) is end-to-end trained in Retrieval Augmented Generation (RAG)?
RAG architecture from the original paper
Since loss is calculated at the output layer of the generator, how the gradients are back propagated to the retriever model?
Because the input to the Generator ...
- 141
0
votes
0
answers
43
views
GAN/DC-GAN isn't converging
I've been trying to train a vanilla GAN(for MNIST) for a few days, and nothing works. I've tried a lot of different layers, hyperparameters, and more, but every time the discriminator's loss decreases(...
- 1
0
votes
0
answers
6
views
Transformation in Normalizing flow
I'm implementing a Flow based model. The affine transformation for decode (z->x) step is: x_{i} = f(x_{i-1}) = x_{i-1}*exp(s) + t, and for encode step is: x_{i-1} = f_1(x_{i}) = (x_{i} - t)/ exp(s)....
- 1
0
votes
1
answer
145
views
Purely extractive Language Model
Given an email thread, I am trying to extract the body of the most recent email.
I used to do that with rules. Now I am testing Large Language Models (LLM) to see if I they provide a less ad hoc ...
- 103
0
votes
0
answers
12
views
Assessing performance of a generative model (images) using Jensen-Shannon Divergence
Background
I am currently working on comparing two high-dimensional image datasets (MNIST) of shape $\sim(50000, 256)$ to evaluate the performance of a generative model, and have attempted to use ...
- 101
1
vote
1
answer
84
views
Unsupervised Machine Translation System Using Variational Autoencoder Models
I want to work on an unsupervised machine translation system using a variational autoencoder. I did a literature review but didn't find any related work, and most of the work is based on denoising ...
0
votes
0
answers
26
views
Improving Normalizing Flows Accuracy
What are some techniques one might use to improve the accuracy of normalizing flows? I am training a flow in a high-dimensional space but it seems like there's always at least one or two dimensions in ...
0
votes
0
answers
25
views
In the GAN objective function, why do we first do we first find the D(x) that maximizes the objective function and then maximise wrt the generator?
The GAN objective function is optimised like this: argmin(argmax(L(G,D))) where the argmax finds the D (Discriminator) that maximises L(G,D). Why is it not the other way around, i.e. argmax(argmin(L(G,...
1
vote
1
answer
2k
views
issue loading the ckpt file PytorchStreamReader failed reading zip archive: failed finding central directory
I am trying to load the ckpt file and getting error
PytorchStreamReader failed reading zip archive: failed finding central directory
Here is the code
...
- 35
2
votes
3
answers
2k
views
torch cuda not able to identify gpu on aws g4dn.xlarge
I have created an EC2 instance with GPU g4dn.xlarge and launched it. I want to run some ...
- 35
1
vote
1
answer
212
views
Why is T5 often used in text-to-data for text prompt encoders?
In the text-to-data(music, image, audio, etc.) generative AI field, one method of encoding text prompts is to use pre-trained language models. Such an approach was used in research on Moûsai [1] and ...
- 33
0
votes
0
answers
11
views
Generating synthetic labeled data (sampling from p(x,y))
I'm working on a toy problem. Consider a dataset that consists of 1-D vectors (waveforms) that contain noise, except for one prominent spike. Denote the waveform by $\vec{x}$, and let the coordinate ...
- 1
1
vote
0
answers
19
views
Avoid overfitting to noise by a noise penalty approach instead of early stopping?
I came across this article on deep learning for computational MRI and found an interesting sentence "However, early stopping has to be performed to not overfit to the noisy measurements." ...
- 11
0
votes
1
answer
55
views
What Deep Learning model to use in this spectroscopy task?
I have a task to be solved. There are energy measurements over the square area 40x40. One measurement consists of values : x, y and the energy. The all area is almost whole covered with data (a few ...
- 101
-1
votes
1
answer
63
views
Build a topic model without data?
I need to come up with a topic model, without any labelled dataset, the model should also be multilingual, thinking of using LLM's as they are accurate and awesome but if Im to build one on my own how ...
- 35
0
votes
0
answers
768
views
How to Use Generative AI for Time Series Forecasting?
What I have
A time series dataset of time stamps (hourly resolution), some covariates (like store foot-traffic) and items sold.
What to forecast
Number of items sold for next 24 hours, i.e. 24 ...
- 335
2
votes
2
answers
167
views
Edit friendly DDPM noise space
I was reading this paper, "An Edit Friendly DDPM Noise Space: Inversion and Manipulations". In page no. 4, they have mentioned that in DDPM, noise maps of consecutive steps are highly ...
- 140
0
votes
0
answers
42
views
Seeking Solutions for Generating Text Descriptions from Diagrams, Infographics, Charts, etc
I am currently on a quest to find an efficient way to generate meaningful text descriptions (or alt text) from visual representations such as diagrams, infographics, charts, plots and the like. ...
- 101