diff --git a/crates/ml/src/trainers/dqn/trainer/training_loop.rs b/crates/ml/src/trainers/dqn/trainer/training_loop.rs
index d03d59cb2..21839d1b5 100644
--- a/crates/ml/src/trainers/dqn/trainer/training_loop.rs
+++ b/crates/ml/src/trainers/dqn/trainer/training_loop.rs
@@ -1247,11 +1247,11 @@ impl DQNTrainer {
             .collect();
 
         self.targets_raw_cuda = Some(
-            crate::cuda_pipeline::clone_htod_f32(&stream, &flat_targets)
+            crate::cuda_pipeline::mapped_pinned::clone_to_device_f32_via_pinned(&stream, &flat_targets)
                 .map_err(|e| anyhow::anyhow!("targets_raw upload: {e}"))?
         );
         self.features_raw_cuda = Some(
-            crate::cuda_pipeline::clone_htod_f32(&stream, &flat_features)
+            crate::cuda_pipeline::mapped_pinned::clone_to_device_f32_via_pinned(&stream, &flat_features)
                 .map_err(|e| anyhow::anyhow!("features_raw upload: {e}"))?
         );
 
diff --git a/docs/dqn-gpu-hot-path-audit.md b/docs/dqn-gpu-hot-path-audit.md
index e43d54abe..356c7e3dc 100644
--- a/docs/dqn-gpu-hot-path-audit.md
+++ b/docs/dqn-gpu-hot-path-audit.md
@@ -160,6 +160,9 @@ Out-of-scope (remaining host-side stream syncs in eval, not part of this fix): `
 - `collect_experiences` (warm but per-collect): `episode_starts_buf` (memcpy_htod) → `upload_i32_via_pinned`; `barrier_config` (htod_f32) → `upload_f32_via_pinned`.
 - `reset_episodes` (warm, per-epoch): replaced the `Vec<f32>` / `Vec<u32>` host staging fields with persistent `MappedF32Buffer` / `MappedU32Buffer` allocated once at construction. Resets fill via `host_slice_mut()` (zero-allocation, zero-HtoD); a single async DtoD seeds the persistent device-resident `portfolio_states` / `rng_states` (which the kernel mutates each step). Adds `host_slice_mut()` accessor on both mapped-pinned types.
 
+### Fix 16 — Migrate `training_loop.rs` raw features/targets uploads (2 sites, 2026-04-28)
+`trainers/dqn/trainer/training_loop.rs:1250` (targets_raw) and `:1254` (features_raw): both `crate::cuda_pipeline::clone_htod_f32` callsites in the data-load helper rewritten to `mapped_pinned::clone_to_device_f32_via_pinned`. These run once per fold immediately before the step loop begins (the `_raw` suffix indicates the un-normalized arrays kept on-GPU for the eval-time critic). The `.map_err(|e| anyhow::anyhow!("…raw upload: {e}"))` wrapping is preserved verbatim — the helper already returns `Result<_, String>` which feeds `anyhow::anyhow!` directly, so the change is a 1:1 path swap.
+
 ### Fix 15 — Migrate `gpu_experience_collector.rs::upload_ofi_features` (1 site, 2026-04-28)
 `gpu_experience_collector.rs:4258`: warm-path `super::clone_htod_f32` upload of OFI features rewritten to `mapped_pinned::clone_to_device_f32_via_pinned`. Called from training_loop's data-load phase before each fold's training run, so the OFI tensor (4M bars × 32 dims = ~512 MB) now stages through mapped-pinned + DtoD just like all other large data uploads. Error-type shift handled with `.map_err(|e| MLError::ModelError(format!("upload_ofi_features: {e}")))?` since the helper returns `Result<_, String>` and the function returns `Result<_, MLError>`.