daniel/h3-blackwell-runtime
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
Direct MiniMax H3 CUDA runtime
90 commits 3 branches 0 tags 34 MiB
Find a file
Daniel Maddern ea7712b4b2 Use paired kitchen RoPE in VAE fast path
2026-08-14 14:58:11 +07:00
benchmarks Initial direct H3 runtime 2026-08-12 14:12:42 +07:00
src/h3_blackwell_runtime Use paired kitchen RoPE in VAE fast path 2026-08-14 14:58:11 +07:00
tools Add VAE decode benchmark tool 2026-08-14 14:36:42 +07:00
wheels Cache GB10 SageAttention3 wheel 2026-08-12 15:21:03 +07:00
.gitignore Initial direct H3 runtime 2026-08-12 14:12:42 +07:00
compose.spark-comfy-lab.yml Initial direct H3 runtime 2026-08-12 14:12:42 +07:00
compose.spark.yml Mount documented H3 artifacts on Spark 2026-08-13 14:44:43 +07:00
Dockerfile.spark Add fastsafetensors loader option 2026-08-13 23:35:01 +07:00
PARITY.md Record sampler wrapper boundary fix 2026-08-13 16:57:44 +07:00
PLAN.md Document H3 audio VAE path 2026-08-14 13:33:08 +07:00
pyproject.toml Add fastsafetensors loader option 2026-08-13 23:35:01 +07:00
README.md Make runtime diagnostics opt in 2026-08-14 14:11:36 +07:00

README.md

H3 Blackwell Runtime

Direct MiniMax H3 Ref2VA runtime research project. ComfyUI is the checkpoint and correctness oracle, not the target runtime.

First Gate

Inspect the mounted H3 NVFP4 safetensors headers before designing an importer:

python .\tools\inspect_safetensors.py /runpod-volume/ComfyUI/models/diffusion_models/minimax_h3_ref2va_pruned_nvfp4.safetensors

Write the output to artifacts/checkpoints/ on the mounted volume. The result must identify packed weights, scales, and tensor naming before any kernel conversion work begins.

Benchmark Contract

benchmarks/ref2va-960x544-124f.json is the single-GPU performance contract. Record direct-runner results as JSON and compare them with:

python .\tools\compare_benchmark.py --result direct-result.json

DGX Spark

Dockerfile.spark and compose.spark.yml prepare an ARM64 GB10 development image using the existing AEON CUDA 13/SageAttention3 base. The compose target opens a shell only; it does not start inference.

Forgejo Pulls From Spark

The Spark checkout uses Forgejo through the host's published local SSH port and a dedicated key:

cd /home/daniel/aeon-spark-test/h3/h3-blackwell-runtime
git config core.sshCommand 'ssh -i ~/.ssh/id_ed25519_forgejo_h3 -o IdentitiesOnly=yes'
git remote set-url origin ssh://git@127.0.0.1:2222/daniel/h3-blackwell-runtime.git
git pull --ff-only origin master

The private key remains on Spark at ~/.ssh/id_ed25519_forgejo_h3; only its public key is registered in Forgejo.

Runtime Output

Generation and latent-decode tools are quiet by default: they suppress ffmpeg banners and only print compact JSON summaries. Use these flags when debugging:

  • --progress: print per-step sampler timing in tools/direct_t2v_preview.py.
  • --profile-memory: print memory checkpoints in tools/direct_t2v_preview.py.
  • --ffmpeg-loglevel info: show ffmpeg details instead of the default error level.
  • --quiet: suppress JSON summary lines.

Standalone tools/compare_*, tools/trace_*, tools/inspect_*, and tools/patch_comfy_* scripts are debugging utilities and remain opt-in by being separate commands.