[not for land] float8 debug recipe for float8 rowwise fwd and hp bwd

[vkuzo]

Summary:

• fwd_rowwise_gi_rowwise_gw_hp - fwd and gi in rowwise, gw in high precision
  • on LLaMa 3 8B, only a 1.7% perf regression vs rowwise_with_gw_hp
• fwd_float8_bwd_hp - fwd in rowwise, bwd in high precision

Test Plan:

with-proxy
CONFIG_FILE="./torchtitan/models/llama3/train_configs/debug_model.toml"
./run_train.sh --model.print_after_conversion --model.converters float8
--training.compile --float8.recipe_name fwd_float8_bwd_hp

with-proxy CONFIG_FILE="./torchtitan/models/llama3/train_configs/llama3_8b.toml" ./run_train.sh --model.print_after_conversion --model.converters float8 --float8.recipe_name fwd_rowwise_gi_rowwise_gw_hp --training.steps 50 --training.compile

Reviewers:

Subscribers:

Tasks:

Tags:

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/2708

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 1 Unrelated Failure

As of commit f0a01fb with merge base 785f3dd ():

NEW FAILURE - The following job has failed:

• PR Label Check / Check PR Labels (gh)
  Process completed with exit code 1.

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• Run 1xH100 Tests / test (H100, linux.aws.h100, --pre torch torchvision torchaudio fbgemm-gpu-genai --index-url https... / linux-job (gh) (trunk failure)
  test/dtypes/test_affine_quantized_float.py::TestAffineQuantizedFloat8Compile::test_expected_kernels_on_gpu_granularity1_torch_compile_mode_reduce-overhead

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[facebook-github-bot]

@vkuzo has imported this pull request. If you are a Meta employee, you can view this in D79777987.