Suffian Khan e6de0eb813 Add nightly pipeline for MI100 to run convergence and batch size test similar to V100. (#6611) ... * Partial updating of ROCM reduction code. * Update reduction_all.cu * Add reduce template parameters. * miopen common * Reuse CUDA's reduction_functions.cc * Reduction ops. * Update remaining reduction ops to use MIOpen. double datatype is not supported, so disable those typed kernels. * Disable a couple more unsupported tests. * Code formatting. * Delete ROCM-specific reduction code that is identical to CUDA reduction code. * Fix scratch buffer early free. * Fix merge conflict. * first attempt nightly amd ci pipeline * try fix bad yaml file * try again with corrected model directory * add convergence test as well * update reference loss for amd mi100 * include mi100 test results csv * update the mi100 convergence test reference values * update batch sizes for mi100 32g * fix gpu sku for run_convergence_test.py * undo unrelated changes to master * pr comments * pr comment Co-authored-by: Jesse Benson <jesseb@microsoft.com>		2021-02-12 13:22:06 -08:00
..
github	Add nightly pipeline for MI100 to run convergence and batch size test similar to V100. (#6611)	2021-02-12 13:22:06 -08:00
__init__.py	Add validation of op registrations (#5817)	2020-11-17 10:44:09 -08:00
amd_hipify.py	Enable more ROCM ops that are sharing CUDA code. Some are needed for Turing NLG models.	2021-02-06 14:40:34 -08:00
build.py	Don't update the excluded ops/types unless args.update is true. Updating the exclusion info triggers rebuilding of all kernels using type reduction. (#6604)	2021-02-09 07:15:31 +10:00
clean_docker_image_cache.py	Fix clean_docker_image_cache.py detection of image pushes. (#6151)	2020-12-16 17:25:22 -08:00
coverage.py	Add support for running Android emulator from build.py on Windows. (#6317)	2021-01-13 19:21:49 -08:00
exclude_unused_ops_and_types.py	Support disabling a typed kernel registration that uses the output type (#6530)	2021-02-03 14:22:32 +10:00
gen_def.py	Reduce IOS shared library size by symbol file. (#5171)	2020-09-14 23:59:41 -07:00
get_docker_image.py	Update get_docker_image.py to enable use without image cache container registry. (#6177)	2020-12-18 19:01:02 -08:00
logger.py	Cache build docker images in container registry. (#5811)	2020-11-17 17:02:24 -08:00
op_registration_utils.py	Support disabling training kernels as part of a reduced build (#6557)	2021-02-09 09:51:31 -08:00
op_registration_validator.py	Support disabling a typed kernel registration that uses the output type (#6530)	2021-02-03 14:22:32 +10:00