Add VAE configuration file

mi300_all_sym_8_step14_fp32 updates.

Add model experiment switching to qkv symmetric quant.

Update unet assets for fused qkv quantization.

Initial commit of FP16 vae model.

Upload updated export irpa.

Update to 20240702 drop from Brevitas.

Update with conv2d bugfix: https://github.com/nod-ai/sharktank/commit/0b67e6aca5185fc517814bd631f821798e1ee26b

Add sharktank exported assets.

Initial add of unet/int8 model.

Update README.md

initial commit