https://github.com/yzhao062/pyod outlier/anomaly detection - many methods in one library #framework #library

https://faker.readthedocs.io/en/stable/index.html
https://sdv.dev/
https://gretel.ai/synthetics

Synthetic Data Generators! #Frameworks #Library

https://medmnist.com/
MedMNIST: A Large-Scale Lightweight Benchmark for 2D and 3D Biomedical Image Classification
#Dataset

https://arxiv.org/pdf/2208.07339.pdf
https://huggingface.co/blog/hf-bitsandbytes-integration
#Performance

https://landing.ai/tips-for-a-data-centric-ai-approach/ #tips Andrew Ng

https://github.com/clovaai/donut #Frameworks #Model Document Understanding Transformer

High-Performance Large-Scale Image Recognition Without Normalization
https://arxiv.org/pdf/2102.06171.pdf #Paper

https://github.com/jsbroks/coco-annotator #Frameworks #Tool Image segmentation tool

https://tf-explain.readthedocs.io/en/latest/index.html
tf-explain offers interpretability methods for Tensorflow 2.0 to ease neural network’s understanding.
#Frameworks

#Tips Efficient Training Large Models on Multiple GPUs, Main Concepts (from https://huggingface.co/docs/transformers/perf_train_gpu_many):

DataParallel (DP) - the same setup is replicated multiple times, and each being fed a slice of the data. The processing is done in parallel and all setups are synchronized at the end of each training step.
TensorParallel (TP) - each tensor is split up into multiple chunks, so instead of having the whole tensor reside on a single gpu, each shard of the tensor resides on its designated gpu. During processing each shard gets processed separately and in parallel on different GPUs and the results are synced at the end of the step. This is what one may call horizontal parallelism, as the splitting happens on horizontal level.
PipelineParallel (PP) - the model is split up vertically (layer-level) across multiple GPUs, so that only one or several layers of the model are places on a single gpu. Each gpu processes in parallel different stages of the pipeline and working on a small chunk of the batch.
Zero Redundancy Optimizer (ZeRO) - Also performs sharding of the tensors somewhat similar to TP, except the whole tensor gets reconstructed in time for a forward or backward computation, therefore the model doesn’t need to be modified. It also supports various offloading techniques to compensate for limited GPU memory.
Sharded DDP - is another name for the foundational ZeRO concept as used by various other implementations of ZeRO.

#Frameworks :
https://www.deepspeed.ai/
https://fairscale.readthedocs.io/en/latest/
https://github.com/tunib-ai/oslo
https://github.com/microsoft/varuna

[2302.14045] Language Is Not All You Need: Aligning Perception with Language Models
https://arxiv.org/abs/2302.14045
#paper New generation

https://github.com/huggingface/peft #Frameworks

https://arxiv.org/abs/2207.06881 #Paper Recurrent Memory Transformer. Scaling transformer architecture to long sequences.

https://www.deepspeed.ai/training/ Large scale training #Frameworks

https://arxiv.org/pdf/2307.02486.pdf Scaling Transformers to
1,000,000,000 Tokens #Paper

https://www.deepmind.com/blog/rt-2-new-model-translates-vision-and-language-into-action
New model translates vision and language into action based on LLM

https://www.anyscale.com/blog/continuous-batching-llm-inference

LLM inference acceleration #Frameworks

https://llava-vl.github.io/blog/2024-01-30-llava-next/

#Frameworks #Models