geohot/tinygrad
1
0
Fork 0
forked from tinygrad/tinygrad
Code Pull Requests Actions Activity

remove cast before view (#8613)

Browse Source 
* remove cast before view

* greener

* indexing

* that passes too

* openpilot too

* ack

---------

Co-authored-by: qazal <[email protected]>
This commit is contained in:
George Hotz
2025-01-14 15:04:58 -05:00
committed by GitHub
co-authored by qazal
parent 393eec3201
commit bfbe81df71
9 changed files with 22 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/benchmark.yml
+3 -3
View File
@@ -91,7 +91,7 @@ jobs:
- name: Run GPT2 w HALF
run: HALF=1 python3.11 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half.txt
- name: Run GPT2 w HALF/BEAM
run: HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 CAST_BEFORE_VIEW=0 python3.11 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
run: HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 python3.11 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
- name: Train MNIST
run: time PYTHONPATH=. TARGET_EVAL_ACC_PCT=96.0 python3.11 examples/beautiful_mnist.py | tee beautiful_mnist.txt
- name: Run 10 CIFAR training steps
@@ -217,7 +217,7 @@ jobs:
- name: Run GPT2 w HALF
run: NV=1 HALF=1 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half.txt
- name: Run GPT2 w HALF/BEAM
run: NV=1 HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 CAST_BEFORE_VIEW=0 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
run: NV=1 HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
- uses: actions/upload-artifact@v4
with:
name: Speed (NVIDIA)
@@ -406,7 +406,7 @@ jobs:
- name: Run GPT2 w HALF
run: AMD=1 HALF=1 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half.txt
- name: Run GPT2 w HALF/BEAM
run: AMD=1 HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 CAST_BEFORE_VIEW=0 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
run: AMD=1 HALF=1 JITBEAM=2 IGNORE_BEAM_CACHE=1 python3 examples/gpt2.py --count 10 --temperature 0 --timing | tee gpt2_half_beam.txt
- uses: actions/upload-artifact@v4
with:
name: Speed (AMD)
.github/workflows/test.yml
+1 -1
View File
@@ -297,7 +297,7 @@ jobs:
- if: ${{ matrix.task == 'optimage' }}
name: Test openpilot model kernel count and gate usage
run: |
PYTHONPATH="." ALLOWED_KERNEL_COUNT=208 ALLOWED_READ_IMAGE=2105 ALLOWED_GATED_READ_IMAGE=29 FLOAT16=0 GPU=1 IMAGE=2 python examples/openpilot/compile3.py https://github.com/commaai/openpilot/raw/v0.9.4/selfdrive/modeld/models/supercombo.onnx
PYTHONPATH="." ALLOWED_KERNEL_COUNT=208 ALLOWED_READ_IMAGE=2104 ALLOWED_GATED_READ_IMAGE=29 FLOAT16=0 GPU=1 IMAGE=2 python examples/openpilot/compile3.py https://github.com/commaai/openpilot/raw/v0.9.4/selfdrive/modeld/models/supercombo.onnx
- if: ${{ matrix.task == 'optimage' }}
name: Test openpilot alt model correctness (float32)
run: PYTHONPATH="." FLOAT16=0 DEBUGCL=1 GPU=1 IMAGE=2 python examples/openpilot/compile3.py https://github.com/commaai/openpilot/raw/3799fe46b3a629e491d4b8498b8ae83e4c88c304/selfdrive/modeld/models/supercombo.onnx