All Questions
Tagged with fine-tuning deep-learning
23
questions
0
votes
0
answers
16
views
RuntimeError: The size of tensor a (128) must match the size of tensor b (122) at non-singleton dimension 2
Description
Error During Fine-Tuning Nvidia TTS Fastpitch Model with Custom Dataset
I am currently trying to fine-tune the FastPitch model from NVIDIA NeMo on a custom dataset but encountered the ...
0
votes
0
answers
41
views
Hyperparameter tuning for detectron2 maskrcnn
For the hyperparameter tuning, below code shows the configuration of the model, i can change its learning rate, iterations, batch size but i am stuck in changing its filter size, activation fuction,...
1
vote
0
answers
114
views
Peft model from checkpoint leading into size missmatch
I have trained peft model and saved it in huggingface. No i want to merge it with base model.
i have used following code.
from peft import PeftModel, PeftConfig,AutoPeftModelForCausalLM
from ...
0
votes
0
answers
20
views
Expected input batch_size (3) to match target batch_size (1)
Context:
Image Captioning model via PyTorch
Fine-Tune with 1 image and a list of captions
Converting the image
Error with prints that show that the size and shape are the same
Before Outputs size: ...
- 21
0
votes
0
answers
301
views
Error while loading MISTRAL LLM for fine-tune. Qlora doesn't work but full works
if I try to load the model in this way :
bnb_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_use_double_quant=True)
model = ...
- 1
0
votes
0
answers
109
views
How can I use the encoder part of the whisper model and sending the output of the encoder to a classification head?
I want to use whisper for speech emotion recognition, and since whisper is an encoder-decoder architecture model, I only want to leverage the encoder part and add a classification head on top of it to ...
0
votes
0
answers
14
views
Why does loss return negative?
`
for _ in range(int(args["num_train_epochs"])):
for step, batch in enumerate(train_dataloader):
model.train()
inputs = {"input_ids": batch[0].to(args["device"]),
"...
0
votes
0
answers
41
views
Which mixed precision setting is in effect when it is set both in accelerate config and TrainingArguments?
I am using both HuggingFace's transformers library and accelerate library.
When it comes to mixed precision setting, there are two places that I can set:
At accelerate config, in the interface below
...
0
votes
0
answers
41
views
Failed to execute (TrainDeepLearningModel) ArcGIS Pro
I was using pre-trained model for human settlement landsat-8 for collecting residential area in ArcGIS Pro. Because the result was not good, i tried to follow this article (https://doc.arcgis.com/en/...
0
votes
0
answers
134
views
Most efficient way to fine tune SDXL for range of product
I want to fine tune SDXL using LoRA on a range of product so that SDXL can generate images of those product later on.
I have many products. What is the most efficient way to fine tune?
Do I just train ...
- 121
1
vote
1
answer
2k
views
Implement Dropout to pretrained Resnet Model in Pytorch
I am trying to implement Dropout to pretrained Resnet Model in Pytorch, and here is my code
feats_list = []
for key, value in model._modules.items():
feats_list.append(value)
for ...
- 15
0
votes
0
answers
403
views
After training the model using SFT, how do I load the model?
I have trained the model with the following code.
from datasets import load_dataset
from trl import SFTTrainer
from transformers import AutoModel, DataCollatorForLanguageModeling, AutoTokenizer, ...
- 1
0
votes
0
answers
295
views
Finetune Wav2vec2 for downstream speech classification
I want to finetune a wav2vec2 model by adding some more custom layers of my own on top of wav2vec for downstream task. Is there any easier way to do this, like just calling the the model without ...
user22329976
1
vote
0
answers
158
views
Low GPU utilization when training PyTorch Model on HPC server, but no issues on personal computer
I'm currently retraining/fine-tuning a visual transformer model (which was pretrained on ImageNet) on Cifar10.
Unfortunately I have problems understanding the performance of the System.
On my personal ...
- 68
-1
votes
1
answer
68
views
Determining the Optimal Approach for Fine-tuning a Pre-trained Neural Network on Images of Varying Sizes
In the context of fine-tuning a pre-trained neural network initially trained on 1024x1024 images, which method is more suitable for adapting a dataset containing images ranging from 320x120 to 320x320?...
