ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
unsloth/scripts
History
Daniel Han 3a21f12500 Studio diffusion (Phase 8): prefer fp8 over nvfp4 in Blackwell auto ladder 
Validated NVFP4 on torch 2.11 + torchao CUTLASS FP4 in an isolated env. The FP4
tensor-core GEMM is genuinely active there (a 16384^3 GEMM hits ~3826 TFLOPS,
2.52x bf16 and 1.37x fp8), but it only beats fp8 on very large GEMMs. At the
diffusion transformer's shapes (hidden ~3072, MLP ~12288, M~4096) NVFP4 is both
slower (0.81x fp8 end to end on Z-Image 1024px) and less accurate (LPIPS 0.166
vs fp8's 0.044). Reorder the Blackwell auto ladder to fp8 before nvfp4 so auto is
correct even on a future MSLK-equipped box; nvfp4 stays an explicit opt-in. Add
scripts/nvfp4_t211_probe.py (extension diagnostics + GEMM micro + end-to-end).
2026-06-26 08:56:23 +00:00
..
data CI: scope GITHUB_TOKEN permissions, add MLX CI, unblock ~60 skipped tests (#5312) 2026-05-11 03:19:13 -07:00
check_frontend_dep_removal.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
check_new_install_scripts.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
compare_engines.py Studio diffusion (Phase 7): max tier uses max-autotune-no-cudagraphs + engine/lever benchmarks 2026-06-26 05:04:10 +00:00
compile_probe.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 03:20:02 +00:00
diffusion_bench.py Studio diffusion (Phase 8): detect consumer vs data-center GPU for fp8 accumulate, with user override 2026-06-26 08:01:15 +00:00
diffusion_quality.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-25 13:56:37 +00:00
enforce_kwargs_spacing.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
fp8_overflow_check.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 07:38:27 +00:00
install_gemma4_mlx.sh Update Install Scripts (#5968) 2026-06-03 05:39:42 -07:00
install_qwen3_6_mlx.sh Update Install Scripts (#5968) 2026-06-03 05:39:42 -07:00
install_rocm_wsl_strixhalo.sh Windows/WSL installer: fix winget msstore cert failure, amd-smi DiskPart prompt, and enable AMD GPU (Strix Halo gfx1151) (#5940) 2026-06-10 04:24:49 -07:00
leverage_probe.py Studio diffusion (Phase 7): max tier uses max-autotune-no-cudagraphs + engine/lever benchmarks 2026-06-26 05:04:10 +00:00
lint_workflow_triggers.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
lockfile_supply_chain_audit.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
notebook_to_python.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
notebook_validator.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
nvfp4_probe.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 08:24:58 +00:00
nvfp4_t211_probe.py Studio diffusion (Phase 8): prefer fp8 over nvfp4 in Blackwell auto ladder 2026-06-26 08:56:23 +00:00
perf_verify.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 03:20:02 +00:00
quant_probe.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 06:17:18 +00:00
run_ruff_format.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
scan_npm_packages.py Package scanners: close fail-open gaps in the sdist fallback and hidden-payload paths (#6359) 2026-06-18 06:50:16 -07:00
scan_npm_packages_baseline.json Package scanners: close fail-open gaps in the sdist fallback and hidden-payload paths (#6359) 2026-06-18 06:50:16 -07:00
scan_packages.py Package scanners: close fail-open gaps in the sdist fallback and hidden-payload paths (#6359) 2026-06-18 06:50:16 -07:00
scan_packages_baseline.json Studio: remove OpenEnv and other unused packages (#6585) 2026-06-23 07:20:47 -07:00
sd_cpp_smoke.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-25 16:08:55 +00:00
sparse_accum_probe.py [pre-commit.ci] auto fixes from pre-commit.com hooks 2026-06-26 06:48:40 +00:00
stamp_studio_release.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
sync_allow_scripts_pins.py Studio: auto-sync allowScripts pins after dependency bumps (#6136) 2026-06-10 02:35:37 -07:00
uninstall.ps1 Windows installer: fix DiskPart UAC mid-install, drive-root cache, and spurious unsloth.exe rename warning (#6296) 2026-06-22 03:09:08 -07:00
uninstall.sh Windows installer: fix DiskPart UAC mid-install, drive-root cache, and spurious unsloth.exe rename warning (#6296) 2026-06-22 03:09:08 -07:00
verify_comment_only_diff.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
verify_import_hoist.py CI: fix import-hoist false positive, vision-cache test cwd, llama.cpp CLI smoke (#6598) 2026-06-23 01:16:47 -07:00