NVIDIA-Merlin/dataloader

Issues

• PyTorch Loader not working

  #160 opened 3 months ago by mohammad-qloo
  2

• [BUG] Unable to extract session embeddings from a session-based transformer model

  #163 opened a year ago by rnyak
  3

• GPU memory does not get freed up properly after each batch

  #108 opened 2 years ago by felix0097
  5

• Out-of-memory error when iterating over merlin.dataloader.torch.Loader

  #158 opened a year ago by richtertill
  1

• [BUG] Data parallel training freezes due to different number of batches

  #75 opened 2 years ago by bschifferer
  17

• Device assignment does not work in PyTorch

  #131 opened 2 years ago by edknv
  4

• [Bug] Conda install command in README cannot find merlin-loader package

  #54 opened 2 years ago by beckernick
  1

• [BUG] Exception in model when using ragged tensors with tensorflow 2.10.0

  #74 opened 2 years ago by oliverholworthy
  9

• Dataloader does not work with tf.keras.layers.Embedding

  #130 opened 2 years ago by edknv
  7

• Feed pre-trained embeddings to NVTabular

  #124 opened 3 years ago by MelissaKR
  6

• [FEA] Data loader: support to padding sparse sequential features on the left side

  #128 opened 3 years ago by gabrielspmoreira
  11

• Shuffle doesn't work

  #85 opened 2 years ago by Ilyushin
  3

• [Feature Request] Make the torch dataloader support TensorDict

  #66 opened 2 years ago by vmoens
  0

• Support aggregation functions used to combine in model item embeddings?

  #33 opened 2 years ago by viswa-nvidia
  0

• Dataloaders-Comparison to petastorm, tensorflow, default pytorch dataloader

  #17 opened 2 years ago by viswa-nvidia
  0

• Change dataloader to output 1D tensors for scalar features

  #105 opened 2 years ago by viswa-nvidia
  1

• Support for 3d tensors (fixed size of sequence) to support session based in TF4Rec and MM

  #34 opened 2 years ago by viswa-nvidia
  1

• NVTabular KerasSequenceLoader costs longer time to load multi-hot features than one-hot features

  #82 opened 2 years ago by Agoniii
  2

• GPU is not detected properly when using SLURM

  #106 opened 2 years ago by felix0097
  5

• [Task] Add multi-gpu data parallel example

  #64 opened 2 years ago by bschifferer
  0

• Typo in `KerasSequenceValidater` class name

  #29 opened 2 years ago by mikemckiernan
  0

• Does this work with images?

  #72 opened 2 years ago by fabiozappo
  1

• [Question] OOM Is there a way not to load the whole dataset in the dataloader?

  #84 opened 2 years ago by gaceladri
  1

• Update Merlin Models to use the new dataloader package

  #14 opened 2 years ago by viswa-nvidia
  0

• Support for images from spark DataFrame

  #56 opened 2 years ago by andompesta
  0

• NVTabular KerasSequenceLoader cost long time to load data

  #46 opened 2 years ago by Agoniii
  3

• Add example for dataloaders for new dataloader repo

  #12 opened 2 years ago by bschifferer
  5

• [BUG] Dataloader doesnt release memory and memory growth

  #76 opened 2 years ago by bschifferer
  6

• Update Transformers4Rec to use new dataloader package

  #16 opened 2 years ago by viswa-nvidia
  2

• Can't import Loader

  #60 opened 2 years ago by dpalbrecht
  4

• How to use dataloader without NVTabular?

  #50 opened 2 years ago by bschifferer
  1

• Add pretrained embedding to the dictionary of tensors

  #32 opened 2 years ago by viswa-nvidia
  0

• Add lookup for embeddings based on key during dataloading

  #31 opened 2 years ago by viswa-nvidia
  0

• Update nvtabular to use the new dataloader package

  #15 opened 2 years ago by viswa-nvidia
  1

• Dataloaders-Docs update

  #20 opened 2 years ago by viswa-nvidia
  2

• Dataloaders-Readme

  #18 opened 2 years ago by viswa-nvidia
  2

• Dataloaders-CI infrastructure

  #19 opened 2 years ago by viswa-nvidia
  0

• Migrate dataloader code from NVT or Merlin Models

  #13 opened 2 years ago by benfred
  1

• Publish new PyPi/Conda packages

  #10 opened 2 years ago by benfred
  0