Update fast TPEC-Quant for proper top_k calculation & lower niter on pca_lowrank

Update README.md

Update README.md

Update fast TPEC-Quant for higher LR cap and --top_k {INT}

Upload experimental fast version of TPEC-Quant

Update README.md

Update README.md

Update README.md

Update TPEC-Quant with faster principal component calcs, add early skips.

Update README.md

Update comment within convert_fp8_scaled_learned_svd.py

Update README.md

Add SVD-based gradient descent scaling script

Add initial fp8 learned converter.

Initial commit