feat: add fp8 training via cublasLt dispatch

- fp8_mm kernel (csrc): cublasLt fp8 e4m3 gemm, TN layout mapped zero-copy
- custom::fp8_mm custom op: meta/cuda/cpu kernels + scale-corrected bf16 autograd
- aten::linear and linear_backward dispatch on CUDA key, zero model changes
- per-tensor scale or raw cast; single-GPU smoke loss matches bf16

astrai/extension/loader.py

@@ -17,6 +17,7 @@ KERNEL_NAMES = [
     "attn_paged_decode",
     "attn_paged_prefill",
     "rotary_emb",
+    "fp8_mm",
 ]
 
 _available: dict[str, bool] = {}