model: batched forward [B,S]

forward_batched(ids[B*S], batch)/loss_batched: run B equal-length sequences as
ONE forward over flattened [B*S] ids, so every linear is one big [B*S,dim] GEMM.
Attention reshapes to [B*nh,S,hd], runs the fused batched causal SDPA (per-seq
mask + RoPE period=S, no cross-sequence attention), writes back [B*S,dim]. The
old per-(batch,head) loop + host-round-tripping split/merge_heads + the additive
causal_mask leaf are gone. forward(ids[seq]) is now forward_batched(ids,1), so
the sampler / inference path (batch=1) is unchanged.

+batched_ids_tensor helper. New batched.rs test: batched forward == looped
single-sequence (logits identical 0.0, grads 6.4e-4, loss identical). PyTorch
parity now exercises B>1 (B=2,S=4): loss 5e-8, logits 6.9e-6, all 25 param
grads within rtol — verifying per-seq RoPE position + per-seq causal masking.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

crates/xtrain-model/src/lib.rs

@@ -24,4 +24,4 @@ pub use config::Config;
 #[cfg(not(no_cuda))]
 mod model;
 #[cfg(not(no_cuda))]
-pub use model::{TinyTransformer, ids_tensor, param_to_host};
+pub use model::{TinyTransformer, batched_ids_tensor, ids_tensor, param_to_host};