Denis Kuznedelev

Problem with the 03_03_vae_digits_train.ipynb

3

Thanks for the book. When running this notebook the program fails on the training step `TypeError: An op outside of the function building code is being passed a "Graph" tensor....

Modified imagenet training script to accept OBSPruning modifier + W&B integration + taking original torchvision models as input. Now one can pass either `OBSPruningModifier` or `MFACPruningModifier` to the ImageNet training...

When running the imagenet training script from torchvision (i.e `src/sparseml/pytorch/image_classification/train.py`) I face the deadlock issue at certain random moment of training (in between the training epoch, usually somewhere in between...

When attempting to run the script ./scripts/finetune_guanaco_7b.sh I face the following error on the beginning of training (after LoRA modules are added): ```bash /opt/conda/conda-bld/pytorch_1682343967769/work/aten/src/ATen/native/cuda/Indexing.cu:1146: indexSelectLargeIndex: block: [30,0,0], thread: [127,0,0] Assertion...

In this pull request following features are added: - full finetuning with teacher logits - validation loss is tracked instead of the rate of loss change on validation - gemma...

Implementation of the RigLPruningModifier - a sparse training procedure proposed in the paper https://arxiv.org/abs/1911.11134. **Description** This sparse training procedure performs sparse training of the model with periodic pruning if a...

The [colab notebook](https://colab.research.google.com/drive/1JQScYICvEC3VqaabLu-lxvq9h7kSV1ML?usp=sharing) with ESRGAN demo has wrong import in the cell `Inference method 2: A minimum snippet of testing codes`. Instead of `from basicsr.models.archs.rrdbnet_arch import RRDBNet` one should use...

Error when installing ffcv from ctime.

3

When trying to install ffcv following the instructions from the repo `README.md`, i.e creation of conda environment followed by pip install, I encounter the following issue when doing `pip install...

Confusing comment in the deeplabv3.yaml

1

In the example of deeplab training in the `yaml` config file deeplab/yamls/deeplabv3.yaml it is written in config file: ```python batch_size: 128 # Training dataloader batch size per device - line...

**Changes** Installation instructions: - Added option of cloning and installing via python `setup.py`. This option is preferrable if one is using `assets` for demo. - Added `diffusers` to requirements. Demo:...