wan22 Configuration files

[christopher5106]

Hi,

I have difficulties to understand how to map all config files in the following folder https://github.com/ModelTC/LightX2V/tree/837feba7d0fe32e4f8ca05024ad125aed4a31ca9/configs/wan22 and the checkpoints I find in https://huggingface.co/lightx2v/Wan2.2-Distill-Models/tree/main

1/ Should I use wan_moe_i2v_distill_quant.json for models: wan2.2_i2v_A14b_high_noise_scaled_fp8_e4m3_lightx2v_4step.safetensors wan2.2_i2v_A14b_low_noise_scaled_fp8_e4m3_lightx2v_4step.safetensors

2/ I see in the file wan_moe_i2v_distill_quant.json "dit_quant_scheme": "fp8-sgl", while in your documentation page I find "dit_quant_scheme": "fp8-vllm" Is it the same ? what is the difference ?

3/ what does _1030 in weight name wan2.2_i2v_A14b_high_noise_scaled_fp8_e4m3_lightx2v_4step_1030.safetensors mean ?

4/ is wan_moe_i2v_distill_quant.json the config file to use for int8 quantisation as well ? (with "dit_quant_scheme": "int8-vllm" addon)

5/ what does _flf2v in wan_distill_moe_flf2v_int8.json mean ? which checkpoints to use with ?