mirror of
https://github.com/unslothai/unsloth.git
synced 2026-07-11 00:38:43 +00:00
fix(studio/rocm): robustness pass - rocm tag normalisation, Strix routing parity, hardened detection
Robustness pass on top of 76137b2d. Four targeted fixes:
1. install.sh ROCm-tag routing normalisation.
`rocm7.2.1` would route to https://download.pytorch.org/whl/rocm7.2.1
which does not exist (PyTorch publishes major.minor URLs only). Same
for any future patch-level tag. Normalise every rocm{maj.min}* pattern
to the bare {maj.min} index URL.
2. install.ps1 + studio/setup.ps1 marketing-name fallback.
The gfx1151 row matched 890M / Strix Halo / HX 37x / HX 38x / AI 9 HX
but not the actual retail name 'AMD Radeon 8060S Graphics' shipped by
OEMs (Ryzen AI MAX+ 395). Add '8060S' to the regex.
3. install_python_stack.py Strix + ROCm 7.1 routing parity with install.sh.
The shell installer reroutes Strix Halo / Point + ROCm 7.1 to
repo.amd.com/rocm/whl/{gfx}/ (which serves torch 2.11.0+rocm7.13.0
with the upstream _grouped_mm fix). The Python `studio update` path
only warned and still installed the broken generic rocm7.1 wheel.
Mirror the override: detect gfx1151/gfx1150 on ROCm 7.1, route to
the AMD per-gfx index, honour UNSLOTH_AMD_ROCM_MIRROR override.
4. _detect_windows_gfx_arch amd-smi parsing tightened.
The amd-smi fallback added in the prior commit used a bare
`\bgfx[1-9][0-9a-z]{2,3}\b` match against the lowercased stdout,
which could pick up stray gfx references in warnings / device-name
strings. Anchor on labelled lines first (Target_Graphics_Version,
ASIC, Arch, gfx) and fall back to the bare match only when no
labelled line is present.
Tests: 231 passed, 1 skipped in tests/studio/install/test_rocm_support.py;
sim_5301 23 cases pass (6 new sims for the Strix override + amd-smi parsing).
This commit is contained in:
Daniel Han
parent
0be974972b
commit
21773215d9
4 changed files with 90 additions and 36 deletions
|
|
@ -1307,7 +1307,7 @@ shell.Run cmd, 0, False
|
|||
$nameArchTable = @(
|
||||
@{ P = "9070 XT|9080"; A = "gfx1201" } # RDNA 4
|
||||
@{ P = "9070|9060"; A = "gfx1200" } # RDNA 4
|
||||
@{ P = "890M|Strix Halo|HX 37[05]|HX 38[05]|AI 9 HX"; A = "gfx1151" } # RDNA 3.5 iGPU (Strix Halo)
|
||||
@{ P = "8060S|890M|Strix Halo|HX 37[05]|HX 38[05]|AI 9 HX"; A = "gfx1151" } # RDNA 3.5 iGPU (Strix Halo / Radeon 8060S retail)
|
||||
@{ P = "880M|Strix Point|AI 9 36[05]|AI 7 35[05]|AI 5 34[05]"; A = "gfx1150" } # RDNA 3.5 iGPU (Strix Point)
|
||||
@{ P = "RX 7900|RX 7800|RX 7700(?! S)"; A = "gfx1100" } # RDNA 3 desktop
|
||||
@{ P = "RX 7600"; A = "gfx1102" } # RDNA 3
|
||||
|
|
|
|||
12
install.sh
12
install.sh
|
|
@ -1644,9 +1644,17 @@ get_torch_index_url() {
|
|||
echo "$_base/cpu"; return ;;
|
||||
esac
|
||||
# Supported tags; 6.5+ clips to rocm6.4, 7.3+ caps to rocm7.2.
|
||||
# PyTorch publishes major.minor URLs only (no patch level), so
|
||||
# rocm7.2.1 / rocm6.0.2 / etc. must normalise to rocm7.2 / rocm6.0.
|
||||
case "$_rocm_tag" in
|
||||
rocm6.0|rocm6.0.*|rocm6.1|rocm6.1.*|rocm6.2|rocm6.2.*|rocm6.3|rocm6.3.*|rocm6.4|rocm6.4.*|rocm7.0|rocm7.0.*|rocm7.1|rocm7.1.*|rocm7.2|rocm7.2.*)
|
||||
echo "$_base/$_rocm_tag" ;;
|
||||
rocm6.0*) echo "$_base/rocm6.0" ;;
|
||||
rocm6.1*) echo "$_base/rocm6.1" ;;
|
||||
rocm6.2*) echo "$_base/rocm6.2" ;;
|
||||
rocm6.3*) echo "$_base/rocm6.3" ;;
|
||||
rocm6.4*) echo "$_base/rocm6.4" ;;
|
||||
rocm7.0*) echo "$_base/rocm7.0" ;;
|
||||
rocm7.1*) echo "$_base/rocm7.1" ;;
|
||||
rocm7.2*) echo "$_base/rocm7.2" ;;
|
||||
rocm6.*)
|
||||
# ROCm 6.5+ (no published PyTorch wheels): clip down
|
||||
# to the last supported 6.x wheel set.
|
||||
|
|
|
|||
|
|
@ -282,10 +282,21 @@ def _detect_windows_gfx_arch() -> str | None:
|
|||
)
|
||||
if result.returncode != 0:
|
||||
continue
|
||||
text = result.stdout.decode(errors = "replace").lower()
|
||||
m = re.search(r"\bgfx[1-9][0-9a-z]{2,3}\b", text)
|
||||
if m:
|
||||
return m.group(0)
|
||||
text = result.stdout.decode(errors = "replace")
|
||||
# Anchor on a labelled gfx line (e.g. "TARGET_GRAPHICS_VERSION: gfx1151"
|
||||
# or "Arch: gfx1151") to avoid catching stray gfx mentions in
|
||||
# warnings or device-name strings. Fall back to a bare token match
|
||||
# only if no labelled line is found.
|
||||
m = re.search(
|
||||
r"(?im)^\s*(?:target_graphics_version|gfx|arch|asic)\b[^:\r\n]*:\s*(gfx[1-9][0-9a-z]{2,3})\b",
|
||||
text,
|
||||
)
|
||||
if not m:
|
||||
m = re.search(r"\bgfx[1-9][0-9a-z]{2,3}\b", text.lower())
|
||||
if m:
|
||||
return m.group(0)
|
||||
continue
|
||||
return m.group(1).lower()
|
||||
except Exception:
|
||||
continue
|
||||
return None
|
||||
|
|
@ -610,42 +621,44 @@ def _ensure_rocm_torch() -> None:
|
|||
|
||||
rocm_torch_ready = has_hip_torch
|
||||
|
||||
# Strix Halo (gfx1151) segfaults under ROCm 7.1 due to a ROCm driver bug
|
||||
# fixed in ROCm 7.2. Warn early so users know why training may crash.
|
||||
# Strix Halo / Strix Point (gfx1151 / gfx1150) segfault under ROCm 7.1
|
||||
# in torch._grouped_mm. AMD's per-gfx repo ships torch 2.11.0+rocm7.13.0
|
||||
# with the real fix, so route those hosts there instead of the generic
|
||||
# pytorch.org rocm7.1 wheel. Mirrors install.sh's Strix override.
|
||||
_strix_override_url: "str | None" = None
|
||||
_strix_override_pkgs: "tuple[str, str, str] | None" = None
|
||||
if ver < (7, 2):
|
||||
gfx_codes = _detect_amd_gfx_codes()
|
||||
_strix_gfx = {"gfx1151", "gfx1150"}
|
||||
if _strix_gfx.intersection(gfx_codes):
|
||||
_gfx_str = ", ".join(sorted(_strix_gfx.intersection(gfx_codes)))
|
||||
_detected_strix = _strix_gfx.intersection(gfx_codes)
|
||||
if _detected_strix:
|
||||
_gfx_str = ", ".join(sorted(_detected_strix))
|
||||
_selected_gfx = sorted(_detected_strix)[0]
|
||||
_amd_mirror = (
|
||||
os.environ.get("UNSLOTH_AMD_ROCM_MIRROR")
|
||||
or "https://repo.amd.com/rocm/whl"
|
||||
).rstrip("/")
|
||||
_strix_override_url = f"{_amd_mirror}/{_selected_gfx}/"
|
||||
_strix_override_pkgs = (
|
||||
"torch>=2.11.0,<2.12.0",
|
||||
"torchvision",
|
||||
"torchaudio",
|
||||
)
|
||||
print(
|
||||
f"\n ⚠️ {_gfx_str} (AMD Strix Halo) detected with ROCm {ver[0]}.{ver[1]}.\n"
|
||||
f" ROCm 7.1 has a known segfault on this GPU when tensors are\n"
|
||||
f" moved to the GPU. Upgrade to ROCm 7.2+ to enable training.\n"
|
||||
f"\n ⚠️ {_gfx_str} (AMD Strix) detected with ROCm {ver[0]}.{ver[1]}.\n"
|
||||
f" ROCm 7.1 has a known _grouped_mm segfault on this GPU;\n"
|
||||
f" routing torch install to AMD's arch-specific index\n"
|
||||
f" ({_strix_override_url}) which serves torch 2.11.0+rocm7.13.0\n"
|
||||
f" with the upstream fix.\n"
|
||||
)
|
||||
|
||||
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]} -- "
|
||||
f"skipping torch reinstall"
|
||||
)
|
||||
else:
|
||||
index_url = f"{_PYTORCH_WHL_BASE}/{tag}"
|
||||
print(f" ROCm {ver[0]}.{ver[1]} -- installing torch from {index_url}")
|
||||
_torch_pkg, _vision_pkg, _audio_pkg = _ROCM_TORCH_PKG_SPECS.get(
|
||||
tag, _ROCM_TORCH_PKG_SPECS["_default"]
|
||||
)
|
||||
if _strix_override_url is not None and _strix_override_pkgs is not None:
|
||||
index_url = _strix_override_url
|
||||
_torch_pkg, _vision_pkg, _audio_pkg = _strix_override_pkgs
|
||||
print(f" Strix ROCm 7.1 override -- installing torch from {index_url}")
|
||||
pip_install(
|
||||
f"ROCm torch ({tag})",
|
||||
"ROCm torch (Strix arch-specific)",
|
||||
"--force-reinstall",
|
||||
"--no-cache-dir",
|
||||
_torch_pkg,
|
||||
|
|
@ -656,6 +669,39 @@ def _ensure_rocm_torch() -> None:
|
|||
constrain = False,
|
||||
)
|
||||
rocm_torch_ready = True
|
||||
else:
|
||||
# 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]} -- "
|
||||
f"skipping torch reinstall"
|
||||
)
|
||||
else:
|
||||
index_url = f"{_PYTORCH_WHL_BASE}/{tag}"
|
||||
print(f" ROCm {ver[0]}.{ver[1]} -- installing torch from {index_url}")
|
||||
_torch_pkg, _vision_pkg, _audio_pkg = _ROCM_TORCH_PKG_SPECS.get(
|
||||
tag, _ROCM_TORCH_PKG_SPECS["_default"]
|
||||
)
|
||||
pip_install(
|
||||
f"ROCm torch ({tag})",
|
||||
"--force-reinstall",
|
||||
"--no-cache-dir",
|
||||
_torch_pkg,
|
||||
_vision_pkg,
|
||||
_audio_pkg,
|
||||
"--index-url",
|
||||
index_url,
|
||||
constrain = False,
|
||||
)
|
||||
rocm_torch_ready = True
|
||||
|
||||
# Install bitsandbytes only when torch links against ROCm. Prefers the
|
||||
# continuous-release_main wheel (bnb PR #1887 4-bit GEMV fix) and falls
|
||||
|
|
|
|||
|
|
@ -792,7 +792,7 @@ if (-not $HasNvidiaSmi) {
|
|||
$nameArchTable = @(
|
||||
@{ P = "9070 XT|9080"; A = "gfx1201" } # RDNA 4
|
||||
@{ P = "9070|9060"; A = "gfx1200" } # RDNA 4
|
||||
@{ P = "890M|Strix Halo|HX 37[05]|HX 38[05]|AI 9 HX"; A = "gfx1151" } # RDNA 3.5 iGPU (Strix Halo)
|
||||
@{ P = "8060S|890M|Strix Halo|HX 37[05]|HX 38[05]|AI 9 HX"; A = "gfx1151" } # RDNA 3.5 iGPU (Strix Halo / Radeon 8060S retail)
|
||||
@{ P = "880M|Strix Point|AI 9 36[05]|AI 7 35[05]|AI 5 34[05]"; A = "gfx1150" } # RDNA 3.5 iGPU (Strix Point)
|
||||
@{ P = "RX 7900|RX 7800|RX 7700(?! S)"; A = "gfx1100" } # RDNA 3 desktop
|
||||
@{ P = "RX 7600"; A = "gfx1102" } # RDNA 3
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue