[R][D]Anaconda env config for LLM research

I'm going crazy trying to get out of dependency hells, and am wondering if anyone has a working conda env they have already setup, would you mind sharing that env config?

Can anyone send a conda env config with the right versions for CUDA(cudatoolkit, nvcc), GCC(and all libraries we need to manually install), TORCH(TORCH-CUDA), TF, DEEPSPEED, TRANSFORMERS and most importantly, Pytorch-Lightning that works without, say DEEPSPEED breaking breaking when used with a Lightning trainer?

anaconda conda cuda gcc install libraries llm machinelearning mind research setup versions