Shifusen

/

Qwen3.5-122B-A10B-abliterated-FP8

like 3

i was run well at Qwen/Qwen3.5-122B-A10B-FP8 but run this model i get an error :
[Shifusen/Qwen3.5-122B-A10B-abliterated-FP8](ValueError: Downcasting not allowed: target.dtype=torch.float8_e4m3fn, loaded_weight.dtype=torch.bfloat16)

i was trying --quantization fp8 then i get another error
ValueError: Quantization method specified in the model config (compressed-tensors) does not match the quantization method specified in the quantization argument (fp8).

GPU : 2 * Sparks ( main + 1 worker )