Stage 7b-i: dense safetensors Qwen3 load path

Adds the bf16/fp16 safetensors path alongside the existing GGUF
quantized one. The harness now dispatches by ModelSpec.quant:
- Some(_) → GGUF (pre-quantized, single-GPU only path, unchanged).
- None    → safetensors dense (new).

The dense path uses candle-transformers::models::qwen3::ModelForCausalLM
verbatim, fed via VarBuilder::from_mmaped_safetensors over the files
listed in `model.safetensors.index.json` (sharded layout) or the
single `model.safetensors` fallback. dtype is bf16 to match the
canonical Qwen3 HF distribution dtype. tokenizer.json is fetched from
the same repo (no -GGUF suffix to strip).

ModelArch gains a Qwen3Dense variant; the forward signature mirrors
QuantizedQwen3Weights (same `forward(&Tensor, offset)` → last-position
logits), so run_inference / run_inference_streaming just add a parallel
match arm — no shape changes downstream.

This is the foundation 7b-ii (ColumnParallel/RowParallel) builds on:
because the source is dense safetensors that can be byte-sliced per
rank, the TP work avoids the GGUF super-block alignment problem
entirely. Vanilla GGUF inference keeps working unchanged.

validate-neuron.sh learns the dense path: pass an empty third arg
(quant) and the script omits the `quant` field from the load
payload, triggering the dense dispatch. Example:
  script/validate-neuron.sh beast.hanzalova.internal Qwen/Qwen3-0.6B ''

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

crates/neuron/src/harness/tp/mod.rs

@@ -213,8 +213,7 @@ impl WorkerPool {
         // Swap out the leader's NcclState into a fresh empty one so we
         // can move it into spawn_blocking; restore after the task
         // returns. (NcclState isn't Clone — it owns a real NCCL Comm.)
-        let mut leader_state =
-            std::mem::take(&mut self.leader_nccl);
+        let mut leader_state = std::mem::take(&mut self.leader_nccl);
         let (returned_state, leader_resp) = tokio::task::spawn_blocking(move || {
             let resp = leader_state.init(leader_cfg, &comm_id_for_leader);
             (leader_state, resp)
@@ -269,8 +268,7 @@ impl WorkerPool {
 
         // 2. Leader's own all_reduce, in spawn_blocking. NCCL operations
         //    block until every rank participates.
-        let mut leader_state =
-            std::mem::take(&mut self.leader_nccl);
+        let mut leader_state = std::mem::take(&mut self.leader_nccl);
         let (returned_state, leader_resp) = tokio::task::spawn_blocking(move || {
             let resp = leader_state.sanity_check();
             (leader_state, resp)