|
Download VALIDATION.md from flashrt/flashrt-flex-attention-train: direct link, hf CLI and curl.
- Browser
- Download file 1.72 kB
-
https://huggingface.co/flashrt/flashrt-flex-attention-train/resolve/main/VALIDATION.md
- Command line
-
hf download hf://flashrt/flashrt-flex-attention-train/VALIDATION.md
-
curl -L -o VALIDATION.md https://huggingface.co/flashrt/flashrt-flex-attention-train/resolve/main/VALIDATION.md
1.72 kB
| # Validation | |
| Correctness smoke: | |
| ```bash | |
| python flashrt-flex-attention-train/tests/test_flashrt_flex_attention_train.py --backend source --mode smoke | |
| python flashrt-flex-attention-train/tests/test_flashrt_flex_attention_train.py --backend source --mode full | |
| ``` | |
| Installed artifact smoke: | |
| ```bash | |
| python flashrt-flex-attention-train/tests/test_flashrt_flex_attention_train.py --backend installed --mode full | |
| python scripts/prebuild_check.py --package flashrt-flex-attention-train --check-config | |
| ``` | |
| Microbenchmark and acceptance gates: | |
| ```bash | |
| python flashrt-flex-attention-train/benchmarks/benchmark.py --device cuda --dtype bf16 --mode all --output /tmp/flex_attention_a100.json | |
| python flashrt-flex-attention-train/benchmarks/benchmark.py --device cuda --dtype bf16 --mode all --require-gates | |
| ``` | |
| Shape/tile matrix: | |
| ```bash | |
| python flashrt-flex-attention-train/benchmarks/shape_matrix.py --presets a100 --output /tmp/a100_flex_matrix.jsonl | |
| python flashrt-flex-attention-train/benchmarks/shape_matrix.py --presets consumer --output /tmp/consumer_flex_matrix.jsonl | |
| ``` | |
| Minimum gate for connecting native kernels: | |
| - forward time <= `0.95 * SDPA` | |
| - forward+backward time <= `0.95 * SDPA` | |
| - peak memory <= `1.03 * SDPA` | |
| - forward max abs diff <= `2e-3` | |
| - gradient norm relative diff <= `1e-2` | |
| Higher gate for publishing native kernels: | |
| - full PI052 step speedup >= `1.08x` on A100 and RTX 5090 | |
| - isolated attention forward+backward speedup >= `1.25x` | |
| - no text/flow step regression | |
| - unsupported shapes automatically fall back to SDPA | |
| The current package provides the reference/fallback path and therefore is not | |
| expected to pass the performance gates until native CUDA fwd/bwd kernels are | |
| added behind the same API. | |