fix binop, other tests failure (#723)

* fix binop, other tests failure

* that was a bad idea

* better layernorm

* inference kernel count tests

* new style reshape pushing

* fixup replacement

* 199 kernels is okay. fix flops

* push reshape through unaryops only

* GRAPH=2 draws the phantom ops

* found resnet issue

* non working test

* mul is cheaper than div

* OPT inflation

* SHUFFLE_PAD_OPS in OPT=2
This commit is contained in:
George Hotz
2023-03-22 18:15:07 -07:00
committed by GitHub
parent d6f4219952
commit b12b60af20
13 changed files with 234 additions and 105 deletions

View File

@@ -165,7 +165,7 @@ jobs:
FORWARD_ONLY=1 GPU=1 IMAGE=2 python3 test/test_ops.py
- name: Test openpilot model
run: |
ALLOWED_KERNEL_COUNT=197 FLOAT16=1 VALIDHACKS=1 DEBUGCL=1 GPU=1 IMAGE=2 python3 openpilot/compile.py
ALLOWED_KERNEL_COUNT=199 FLOAT16=1 VALIDHACKS=1 DEBUGCL=1 GPU=1 IMAGE=2 python3 openpilot/compile.py
python3 -c 'import os; assert os.path.getsize("/tmp/output.thneed") < 100_000_000'
DEBUGCL=1 GPU=1 IMAGE=2 python3 openpilot/compile.py
VALIDHACKS=1 DEBUGCL=1 GPU=1 IMAGE=2 python3 openpilot/compile.py