Skip to main content

Questions tagged [training]

For questions about training networks, rules systems, or other AI system components.

Filter by
Sorted by
Tagged with
0 votes
0 answers
6 views

How to 'compose' a sport video based on .. action

I'm recording some football game using two "fixed" action cam, starting recording both at the same time. I'm here to ask you the best approach (if exists) to train a program to take each ...
stighy's user avatar
  • 101
-1 votes
0 answers
15 views

The training plots are different between the original PID controller and the behavioral cloner (LSTM) of this controller

Here is the code i used however by looking at the figures it so weird and i can't figure out the problem: import torch import torch.nn as nn import torch.optim as optim from scipy.integrate import ...
Sarette's user avatar
2 votes
1 answer
18 views

How does Training and Validating work with Graph CNNs

I'm training a Graph Convolutional Neural Network to output embeddings for nodes that I eventually want to perform classification on. I am a little confused on how the training, validation, and ...
Kiran Manicka's user avatar
0 votes
0 answers
29 views

Model suggestion for AI based scaling

We are exploring the idea of scaling elements within a UI container based on the given size. The container is represented by a json object, for example: ...
Sameed's user avatar
  • 101
0 votes
0 answers
13 views

How to Implement a Text-Based Question Answering System for PDF Documents using Machine Learning?

As a beginner in machine learning, I've completed a basic text classification project in university. Now, I'm eager to build a system that can answer specific questions from a large collection of PDF ...
Rudra Sarkar's user avatar
0 votes
0 answers
8 views

Can you iteratively freeze and unfreeze parts of a neural network for efficient training?

I know you can do efficient training by freezing parts of a NN, but is there any work done where part 1 of a NN is frozen and part 2 is trained, and then part 2 is frozen and part 1 is trained?
JobHunter69's user avatar
0 votes
0 answers
15 views

How are conditional models different from supervised models?

I'm wondering what the difference between conditional learning and supervised learning is - especially in diffusion models? Am I correct to assume that diffusion models are supervised because in ...
euleriwt's user avatar
0 votes
0 answers
11 views

From how many experts does LLM training using a mixture of experts (MoE) start slowing down compared to training LLMs without MoE?

I'm trying to find some information regarding the impact of the number of experts on LLM training. From how many experts does LLM training using a mixture of experts (MoE) start slowing down compared ...
Franck Dernoncourt's user avatar
11 votes
2 answers
6k views

Can you train a neural network by simply giving it ratings each time it runs?

I am currently trying to train a bot for a game I am creating. It is a 2d game with a complex map made of various shapes. The bot and character shoot bullets that are capable of ricocheting. The ...
Beluker's user avatar
  • 113
0 votes
0 answers
19 views

Flipping train and test labels for binary classification

I was training a GCN (this one) on a single graph (n=1,1304 nodes, num_features=26) to perform node level binary classification. However, my model performed with 5% accuracy (and even went as low to 0%...
Heisen _'s user avatar
0 votes
0 answers
17 views

Multimodal architecture: how to train it

I have a set of graphs, a set of sequences and a set of summary stats, the elements in the three sets correspond to each other. I am not satisfied with the regression performance from one single ...
Tianjian Qin's user avatar
-1 votes
0 answers
21 views

Error in finetune paddleOCR

Anyone help me this error in finetuning paddleOCR. Here the link to Notebook https://www.kaggle.com/code/ryanlliu/finetune-paddleocr?scriptVersionId=169451466 and dataset https://www.kaggle.com/...
Alex Luu's user avatar
0 votes
1 answer
37 views

model.fit fails using keras sequential (slice index ### of dimension 0 out of bounds) [closed]

this is the most simple model I can think of for my data yet I can't use the fit function, it gives an error. the desired procedure is to make a simple autoencoder : from 576 nodes to 64 then back to ...
Bikay's user avatar
  • 23
2 votes
1 answer
61 views

What are these special "AI chips" actually used for?

To make an AI / LLM, like ChatGPT, you need two things: To create the LLM. This includes training it, etc. Very expensive from computation perspective. Run the LLM to answer user queries. For ...
ineedahero's user avatar
0 votes
0 answers
21 views

Can we use quantization techniques to train ML models from scratch?

I'm currently learning about quantization and have seen that most methods for quantizing are for a Post-Training conversion or, in some cases, for efficient finetuning like the QLORA method. I'd like ...
Cesar Ruiz's user avatar
0 votes
2 answers
74 views

Why are hallucinations a huge problem?

Why it is not easy to just train model for factual accuracy and punish for any invented facts or mistakes, award for admitting ignorance in difficult cases?
Anixx's user avatar
  • 331
1 vote
0 answers
61 views

AI chat bot that answers by focusing only on 30 textbooks [closed]

I don't even know what I'm looking for and what's the terminology, so here I am asking this question. Background Assume I have 30 textbooks. I want to have an AI chatbot like ChatGPT which answers the ...
Megidd's user avatar
  • 141
0 votes
0 answers
15 views

What kind of model is needed to train generative avatar with available video training material?

Joining this community, first post for me, happy to be here. As a personal exercise, I'd like to develop an app similar to the Biden vs Trump endless debate https://www.twitch.tv/trumporbiden2024 just ...
TMOTTM's user avatar
  • 101
0 votes
0 answers
38 views

Does DPO update the weights in the same way that LORA/Fine-Tuining does?

I've been working on making a conversational customer service fine-tune for the past couple months and now I am looking to improve its failure recovery. For example, If the agent makes a mistake, how ...
the-test-set-is-all-you-need's user avatar
0 votes
1 answer
63 views

How LSTM really decide what to forget and not?

Currently, I am learning about LSTM, and I understand the intuition behind it, such as how forget gate works (sigmoid function yields a value between 0 and 1; if it is 0 it "completely" ...
Ashraf's user avatar
  • 3
0 votes
0 answers
13 views

adversarial training on convnext shows a very strange curve

i am currently working on a research project where I have to train some models for adversarial robustness. I have implemented the algorithm used by a research paper called adversarial training for ...
M Akrm's user avatar
  • 1
0 votes
0 answers
41 views

The end-to-end Training Process for Knowledge Distillation

I'm a bit confused on the complete training process for Knowledge Distillation. I was reading the Geoffrey Hinton "Distilling the Knowledge in a Neural Network" 2015 paper and some random ...
Chuu's user avatar
  • 1
0 votes
1 answer
79 views

What does it mean if I trained my model with more steps per epoch than the total number of training images I have?

I'm having a little bit of trouble understanding what steps per epoch really means. I've read that Number of Steps per Epoch = (Total Number of Training Samples) / (Batch Size), however I don't ...
Triana Anderson's user avatar
0 votes
0 answers
12 views

3-way hold-out for picking between different ML models

I have a dataset which is a time series. Before jumping on more heavy models such as LSTM, I wanted to test out the performance of linear models. I have currently a 80/20 split between training and ...
xingern's user avatar
0 votes
1 answer
373 views

Getting started with training local LLM using python [closed]

As I'm completely new to this field, I find it hard to get started given the requirements I have. I'm a bit overwhelmed by all the models and options that are available. Even though it wasn't ...
Jeanluca Scaljeri's user avatar
0 votes
0 answers
16 views

I need some direction for a solution on querying gigabytes of text

I have a bunch of old automotive magazines from the 1900's all the way up to the 1990's with (i imagine) some very old and lost forgotten info. The magazines have been converted to PDF, OCR'd and then ...
Vince's user avatar
  • 1
2 votes
1 answer
212 views

Human trainers: collect data by 2-alternative force choice (2AFC)

To collect human intelligence data, we intend to use 2-alternative force choice (2AFC). A sample human intelligence task (HIT) is shown below. Can it be considered a 2AFC? Is such a HIT considered ...
Megidd's user avatar
  • 141
0 votes
1 answer
113 views

Does Chat-GPT learn from interactions? [closed]

Does Chat-GPT accumulate knowledge through interactions with users in real time? I know it's trained on a massive dataset but does it do online learning?
FourierFlux's user avatar
0 votes
1 answer
35 views

Validation loss is always lower than training loss whatever i try

I've been training several types of MLPs with different optimisers and tuned them with keras's hyperband tuner. All of them follow this cone architecture: All the networks were trained on the same ...
Roger Smith's user avatar
0 votes
0 answers
19 views

Optimal number of epochs for Transformer network on time series data?

I have a transformer network that is trained on time series data. The task is to predict if a variable will increase a certain percentage in the next dt days. The input is data from the 90 previous ...
QCQCQC's user avatar
  • 101
1 vote
0 answers
555 views

Finetuning Mistral or MistralForSequenceClassification for text classification

I need to do text classification and have a dataset of 10K entries. I am considering using mistral and following a tutorial like https://huggingface.co/docs/transformers/training and replace model ...
Karl 17302's user avatar
0 votes
0 answers
19 views

If you throw additional compute at DQN will you get the same results as if you threw compute at transformers?

Recent performance in transformer based AI models have done really well by just throwing additional compute and data into the training. Can you expect similar levels of results by throwing more ...
Dylan Kerler's user avatar
1 vote
1 answer
338 views

What is accelerated years in describing the amount of the training time?

As described in this article, it was written that GPT-3 took 405 V100 years to train in 2020. I'm a bit confused about this definition, does that mean the process was accelerated like using a V100 GPU ...
Marziye Hasanshahi's user avatar
1 vote
0 answers
52 views

Largest fully open source LLMs including training

It seems that most of the open-sourced large language models (LLMs) like Llama 2 had the model released but not the exact training procedure and training data-sources (exact data revisions) so that ...
Kozuch's user avatar
  • 281
1 vote
1 answer
498 views

Which epoch is the best for me to choose?

I have trained my deep learning model. I also saved the validation loss to a file and plotted on a graph I have $2$ questions for this: Does the validation loss look normal? Is there any issue with ...
user avatar
1 vote
3 answers
70 views

Would maximizing (instead of minimizing) error of an LLM/HMM lead to complex behavior?

Imagine we have some sort of "next token predictor," either with transformer architecture, LSTM, or just a HMM (though the terminology I use here will be less aligned to HMMs, I believe the ...
BigMistake's user avatar
1 vote
0 answers
37 views

Neural network learns to mimic distribution of classes in dataset instead of using signal from input

I'm trying to implement example from a classic AI paper named "Learning representations by back-propagating errors" by Hinton et al. Example aims at training network able to predict third ...
Jan Grzybek's user avatar
0 votes
3 answers
113 views

My model is only improving when learning rate is 1. Should I be worried?

As the title says my GNN with three layers of GAT (Graph attention layers) is only moving the metrics when the learning rate is 1. As generally the learning rate is (0,1) should I be worried? Also ...
DataDoge's user avatar
0 votes
0 answers
31 views

How was the word2vec model trained?

Let's take the CBOW (continuous bag of words) model as the example. Suppose that, there are $c$ context words, each of which is a one-hot encoding vector. So the total number of elements of input ...
J. Doe's user avatar
  • 1
0 votes
0 answers
11 views

Classes definition for detecting impervious surfaces on aerial photographies

My project is to use deep learning, essentially a UNET segmentation model, to detect impervious surfaces on high resolution aerial photographies. I wonder if it's better to train the model with many ...
Below the Radar's user avatar
0 votes
0 answers
40 views

From Text-To-Speech to LLMs: Providing "writing style"

I've just recently learned about Text-To-Speech models and how they are trained. Unlike LLMs, to a provided pair (text, speech), a feature vector ${f}$, that was generated by more speech of that ...
Mathy's user avatar
  • 143
0 votes
0 answers
28 views

use AI to creat images from texts

I want to use AI to create art images from a single text available in different languages, with the generated images being different depending on the language of the given text. example 1: step 1: the ...
Svirepenisch's user avatar
1 vote
0 answers
35 views

How do LGBM rankers train?

I'm looking into Learning to Rank models - specifically, the LGBMRanker model - and I want to understand how it's able to train. It takes in features, group sizes and labels, and optimizes for a ...
Shirish Kulhari's user avatar
2 votes
1 answer
160 views

How to train a sample weight model for another ML model?

I'm trying to train a ML model, however the predictability of the different samples varies, i.e. some samples are inherently much harder to predict/estimate than others. Poorer predictions for these ...
Hiho's user avatar
  • 123
0 votes
0 answers
37 views

Multi-Agent DQN not learning for Clean Up Game - Reward slowly decreasing

The environment of the Clean Up game is simple: in a 25*18 grid world, there's dirt spawning on the left side and apples spawning on the other. Agents get a +1 reward for eating an apple (by stepping ...
Charles's user avatar
0 votes
0 answers
22 views

High Accuracy ML.NET Image differentiation model

I have a relatively big dataset (100+GB) that has 35 categories. All of them are microscopic images with slight differences. Although ML.NET documentation itself declares that training time should be ...
Helios Lucifer's user avatar
1 vote
2 answers
400 views

Tips and tricks when training a very large language model?

Have never trained a (very) large language model, so I am wondering if the process is the same as training a (regular) language model, i.e. you prepare the data, set up the architecture, ...
information_interchange's user avatar
2 votes
1 answer
836 views

Why do transformers compute the loss over the prompt?

When fine-tuning large language models, which are commonly decoder transformer architectures, sometimes we want to compute the training loss over the entire prompt, sometimes just the completion ...
milez's user avatar
  • 981
0 votes
1 answer
42 views

How to label missing/default values for a named entity recognition dataset

I am building the training dataset for a named entity recognition model, with 2 tags: Name and Category and I am using a pre-...
mrang's user avatar
  • 3
0 votes
0 answers
9 views

Mechanism of Prediction Readjustment in Supervised Learning and Role of Self-Attention in Sequence Data Relationships

In supervised learning, when the prediction deviates significantly from the expectation, how does it "readjust"? And... LLMs are a subset of deep learning, just as generative AIs are. Is the ...
Assandra Lakal's user avatar

1
2 3 4 5
10