in this place, i store some library that i think it will help everyone in working process
- loguru: simple but flexible logging
- wandb: tracking experiments
- tqdm: loading bar
- t-sne cuda: accelerate t-sne algorithm with cuda (instead of sckit-learn tsne)
- vllm: accelerate inference with saved memory
- unsloth: accelerate fine-tuning LLM (save, faster) (multiGPU support coming soon)
