Issues
- 1
Library versions of Python Dependencies
#158 opened by mostafaelhoushi - 0
How to get peak memory of solution
#157 opened by haoming-codes - 0
- 3
Question about profile-based cost model
#154 opened by jasperzhong - 1
Can I use it on a custom training loop?
#155 opened by hermesonbf - 3
Checkmate solver fails while the original model can be run with a specific memory budget
#151 opened by hy00nc - 1
Transformer benchmark
#153 opened by eric-haibin-lin - 3
Meta-issue tracking PyTorch integration attempts
#126 opened by parasj - 4
edit_graph function
#150 opened by supersoob - 1
- 0
- 0
is_gpu_available is deprecated
#148 opened by parasj - 0
Add full support for Python 3.8 into build matrix
#149 opened by parasj - 0
Add `schedule_save_path` and `schedule_load_path` arguments to the compile_tf2 function
#131 opened by parasj - 0
- 0
- 0
Add profiling to wrapper.py
#134 opened by parasj - 0
[performance] Calculating APs is slow
#109 opened by parasj - 0
[performance] copy_op in execution.py is very slow
#110 opened by parasj - 1
Bug - installing via pip does not work
#113 opened by parasj - 1
- 0
- 1
Publish checkmate to PyPI
#100 opened by parasj - 0
- 2
Document Checkmate codebase
#83 opened by parasj - 0
Checkmate is not compatable with Python 3.5
#88 opened by parasj - 0
Checkmate is not compatible with Python 2.7
#93 opened by parasj - 0
- 0
- 0
Add tests for LP approximation strategies
#76 opened by parasj - 9
compile_tf2 fails in the tutorial
#144 opened by sarangsable - 1
Collating power numbers for Maker1000
#135 opened by parasj - 2
Add documentation to install extras.
#115 opened by aninrusimha - 1
[BUG] Griewank should adapt to non-linear graphs
#69 opened by parasj - 1
Profiling
#55 opened by aninrusimha - 2
Run TF Execution on Mobilenet, resnet
#57 opened by aninrusimha - 1
- 1
- 0
Create a dfgraph with constant ops per layer [can't do flop counts yet, lower priority than profiling]
#56 opened by aninrusimha - 2
- 2
- 1
Squash the NLP code as a new TestBERT network
#59 opened by parasj - 1
- 0
Experiments should make a new timestampped directory rather than deleting the contents of an existing experiment data directory
#43 opened by parasj - 0
- 1
Add makefile to format code, run tests, install packages (gpu versus non-gpu) etc.
#38 opened by parasj - 1
[refactor] Cache ILP results in Redis
#19 opened by parasj - 1
Budget sweep is broken
#74 opened by parasj - 0
- 1