Fix round 2 regressions: ROCm validate_server and Windows HIP routing

Follow-up to 810b833b addressing review findings on the first round of
hardening commits:

- install_llama_prebuilt.py validate_server: gate --n-gpu-layers on the
  resolved install_kind instead of host.has_rocm. AMD Windows hosts
  without a HIP prebuilt fall back to windows-cpu and must not be
  validated with GPU layers; thread install_kind through from the
  caller.
- install_llama_prebuilt.py resolve_release_asset_choice: reinstate the
  "not has_rocm" guard on the published windows-cpu bundle so AMD
  Windows hosts reach resolve_upstream_asset_choice() where the new
  HIP prebuilt path lives. Prefer a published windows-hip bundle first
  when one exists, fall through to upstream HIP + upstream CPU
  otherwise.
- install_llama_prebuilt.py detect_host: also set has_physical_nvidia
  when the secondary --query-gpu block confirms a working NVIDIA GPU,
  so older nvidia-smi versions without -L support do not silently skip
  the Linux diagnostics that key off has_physical_nvidia.
- install_llama_prebuilt.py: drop redundant "import re as _re" /
  "import re as _re_rocm" local aliases in favour of the existing
  top-level "import re".
- install_python_stack.py _ensure_rocm_torch: run the AMD
  bitsandbytes install unconditionally after the HIP-torch probe so
  "unsloth studio update" on venvs that already have ROCm torch still
  gains the AMD bitsandbytes build.
- install.sh: add a non-x86_64 early-exit to get_torch_index_url() so
  aarch64 / arm64 Linux hosts do not hit the ROCm wheel index
  (PyTorch only publishes ROCm wheels for linux_x86_64).
- install.sh: add bitsandbytes install to the migrated-environment
  branch so upgrades pick it up for ROCm hosts instead of only the
  fresh-install path.
- install.sh: in the Radeon wheel path, pass version constraints +
  --no-index --find-links to uv instead of explicit wheel URLs so a
  version-compatible torch / torchvision / torchaudio triple is
  resolved, rather than picking the highest-version wheel for each
  package independently.
- studio/backend/utils/hardware/amd.py _first_visible_amd_gpu_id: fall
  through to lower-priority visibility env vars when the first entry
  is malformed (leading comma, all-whitespace first token) instead of
  silently returning GPU 0.

studio/install_python_stack.py

@@ -185,9 +185,9 @@ def _ensure_rocm_torch() -> None:
         print("   ROCm detected but version unreadable -- skipping torch reinstall")
         return
 
-    # Skip if torch already links against HIP (ROCm is already working).
-    # Do NOT skip for CUDA-only builds since they are unusable on AMD-only hosts
-    # (the NVIDIA check above already handled mixed AMD+NVIDIA setups).
+    # Probe whether torch already links against HIP (ROCm is already working).
+    # Do NOT skip for CUDA-only builds since they are unusable on AMD-only
+    # hosts (the NVIDIA check above already handled mixed AMD+NVIDIA setups).
     try:
         probe = subprocess.run(
             [
@@ -201,36 +201,43 @@ def _ensure_rocm_torch() -> None:
         )
     except (OSError, subprocess.TimeoutExpired):
         probe = None
-    if probe is not None and probe.returncode == 0 and probe.stdout.decode().strip():
-        return  # torch already has HIP/ROCm backend
-
-    # Select best matching wheel tag (newest ROCm version <= installed)
-    tag = next(
-        (
-            t
-            for (maj, mn), t in sorted(_ROCM_TORCH_INDEX.items(), reverse = True)
-            if ver >= (maj, mn)
-        ),
-        None,
+    has_hip_torch = (
+        probe is not None
+        and probe.returncode == 0
+        and probe.stdout.decode().strip() != ""
     )
-    if tag is None:
-        print(f"   No PyTorch wheel for ROCm {ver[0]}.{ver[1]} -- skipping")
-        return
 
-    index_url = f"{_PYTORCH_WHL_BASE}/{tag}"
-    print(f"   ROCm {ver[0]}.{ver[1]} -- installing torch from {index_url}")
-    pip_install(
-        f"ROCm torch ({tag})",
-        "--force-reinstall",
-        "--no-cache-dir",
-        "torch>=2.4,<2.11.0",
-        "torchvision<0.26.0",
-        "torchaudio<2.11.0",
-        "--index-url",
-        index_url,
-        constrain = False,
-    )
-    # Also install bitsandbytes for AMD
+    if not has_hip_torch:
+        # Select best matching wheel tag (newest ROCm version <= installed)
+        tag = next(
+            (
+                t
+                for (maj, mn), t in sorted(_ROCM_TORCH_INDEX.items(), reverse = True)
+                if ver >= (maj, mn)
+            ),
+            None,
+        )
+        if tag is None:
+            print(f"   No PyTorch wheel for ROCm {ver[0]}.{ver[1]} -- skipping")
+            return
+
+        index_url = f"{_PYTORCH_WHL_BASE}/{tag}"
+        print(f"   ROCm {ver[0]}.{ver[1]} -- installing torch from {index_url}")
+        pip_install(
+            f"ROCm torch ({tag})",
+            "--force-reinstall",
+            "--no-cache-dir",
+            "torch>=2.4,<2.11.0",
+            "torchvision<0.26.0",
+            "torchaudio<2.11.0",
+            "--index-url",
+            index_url,
+            constrain = False,
+        )
+
+    # Always install bitsandbytes for AMD -- runs even when torch was not
+    # reinstalled (e.g. "unsloth studio update" on a venv that already has
+    # ROCm torch) so the AMD bitsandbytes dependency is not left missing.
     pip_install(
         "bitsandbytes (AMD)",
         "--no-cache-dir",