Megvii Engine Team
e918f0aa75
feat(traced_module): add treedef leaf node check and add some graph api
GitOrigin-RevId: 36c069bfee
3 years ago
Megvii Engine Team
c7e730bc12
feat(traced_module): add some functions of graph modification
GitOrigin-RevId: 09691ebd33
3 years ago
Megvii Engine Team
f88bd3ae33
refactor(traced_module): let TracedModule own argdef_graph_map
GitOrigin-RevId: 80d685b9a3
3 years ago
Megvii Engine Team
b1c46ba46c
feat(traced_module): add some functions of graph modification
GitOrigin-RevId: ac0603057a
3 years ago
Megvii Engine Team
4bb253695b
feat(traced_module): let CallFunction own graph
GitOrigin-RevId: 66cdbca7e5
3 years ago
Megvii Engine Team
9a6a379346
feat(traced_module): add visit method
GitOrigin-RevId: 251ecebf87
4 years ago
Megvii Engine Team
442b4f6c26
test(traced_module): add some testcases for traced module
GitOrigin-RevId: 0d6bb20b2b
4 years ago
Megvii Engine Team
f2691566fd
feat(traced_module): add pytree
GitOrigin-RevId: 6c6e53521c
4 years ago
Megvii Engine Team
bee305beb2
feat(traced_module): add functional trace and CallMethod/Function expr
GitOrigin-RevId: ad2cdc1b61
4 years ago
Megvii Engine Team
763c56f3b9
feat(imperative): add traced module
GitOrigin-RevId: 28c3503f2e
4 years ago
Megvii Engine Team
9279104b11
feat(mge): add opdef serialization and apply_module_trace
GitOrigin-RevId: 5b45bded1d
3 years ago
Megvii Engine Team
aa20404027
feat(lite): add lite static all in one
GitOrigin-RevId: 7e6d15d929
3 years ago
Megvii Engine Team
a0231a7920
fix(dnn/cuda): fix algo matmul for conv bwd filter
fix fastrun workspace size not available exception and device OOM error caused by the incorrect workspace size calculation of algo matmul of conv bwd filter
GitOrigin-RevId: de96b4fe11
3 years ago
Megvii Engine Team
f3ed59d336
feat(dnn/opencl): add heuristic rule for elemwise
GitOrigin-RevId: 2bc574b5a7
3 years ago
Megvii Engine Team
29d24dbb80
fix(mge/function): fix interpolate unsupport fp16 error
GitOrigin-RevId: 7fc6271986
3 years ago
Megvii Engine Team
36df3850f3
test(mgb): remove the padding random test case
padding random test case in ci 2080ti will failed in random
GitOrigin-RevId: b83e7b8799
3 years ago
Megvii Engine Team
e21967bb40
feat(mge): add env MGE_FASTRUN_CACHE_DIR
GitOrigin-RevId: 0351ff88d1
3 years ago
Megvii Engine Team
6a1ec8a890
feat(mge): add git commit-id into fastrun cache key
GitOrigin-RevId: fb614b4d9c
3 years ago
Megvii Engine Team
ae87876d34
feat(mge): refactor weightscaler
GitOrigin-RevId: 7f874388f7
3 years ago
Megvii Engine Team
5d9ac970ab
fix(mgb): fix fastrun compnode
GitOrigin-RevId: 8db93facb9
3 years ago
Megvii Engine Team
56c1b626bf
refactor(dnn): move arch-dependant code to arch.h
GitOrigin-RevId: 52350144b1
3 years ago
Megvii Engine Team
67575d582c
feat(mge/opr): add interpolate bilinear mode
GitOrigin-RevId: f7023a3fd3
3 years ago
Megvii Engine Team
0558b2123d
feat(mge/opr): add interpolate nearest mode
GitOrigin-RevId: d384b87f50
3 years ago
Megvii Engine Team
171d69155a
fix(fp16): fix midout build issue when hit fp16 trace
GitOrigin-RevId: cf2c5184cd
3 years ago
Megvii Engine Team
127870a926
feat(dnn/opencl): add heuristic rule for batched matmul
GitOrigin-RevId: bd152428e6
3 years ago
Megvii Engine Team
d86ed426ee
fix(dtr): simulate the system stack to avoid stack overflow during recomputing
GitOrigin-RevId: cb73e62b19
3 years ago
Megvii Engine Team
c25125e3d2
perf(dnn/cuda): sass int8 epilogue remove shared load
GitOrigin-RevId: 2b49f5069b
3 years ago
Megvii Engine Team
bc2b1690c9
ci(thirdparty): add third_party cache
GitOrigin-RevId: d54681f0c0
3 years ago
Megvii Engine Team
6070f1272d
fix(mgb): fix getting static memory alloc info
GitOrigin-RevId: dfc69c3b3f
4 years ago
Megvii Engine Team
e8a5932d1e
perf(mgb/gopt): optimize impl of reformat builders
GitOrigin-RevId: 844b7e8d39
3 years ago
Megvii Engine Team
58b8b14554
refactor(mgb/gopt): add checker for reformat emitter
GitOrigin-RevId: 53a8c128f5
3 years ago
Megvii Engine Team
55efc8e197
feat(mgb/gopt): add reformat emitter
GitOrigin-RevId: 937b20a57c
4 years ago
Megvii Engine Team
c9d060307f
feat(dnn/common): add named tensor shape
GitOrigin-RevId: 918928b8ba
4 years ago
Megvii Engine Team
ff0e6be7b9
fix(dnn/cuda): fix cutlass tensorop kernels
do not compile cutlass tensorop kernels, when using cuda version less than 10.2
GitOrigin-RevId: d4c37d5f41
3 years ago
Megvii Engine Team
336761253d
feat(dnn/cuda): add tensorcore matmul for fp16 data type
GitOrigin-RevId: 025c591f75
3 years ago
Megvii Engine Team
12cdbddd14
fix(ci): clean fastrun cache in windows and macos ci
GitOrigin-RevId: d1a010287f
3 years ago
Megvii Engine Team
31705913c0
fix(ci): set MGE_FASTRUN_CACHE_TYPE=FILE in ci env
GitOrigin-RevId: c4a549480e
3 years ago
huangxinda
f814a4ae78
ci(mge): update test script
3 years ago
Megvii Engine Team
2c4ee99227
fix(dnn): short cutlass filename in windows
GitOrigin-RevId: 83a43fdf87
3 years ago
Megvii Engine Team
b17b56f309
fix(build): fix copy bara error
GitOrigin-RevId: 6d68824821
3 years ago
Megvii Engine Team
3c6665f7c1
feat(lite/whl): merge lite whl to main package
GitOrigin-RevId: 27c7e50207
3 years ago
Megvii Engine Team
989fdde255
refactor(subgraph): use graph queue to cache compiled op graphs
GitOrigin-RevId: cba8574c73
3 years ago
Megvii Engine Team
a7a3bf2d6c
test(subgraph): simple test for subgraph
GitOrigin-RevId: 3d6ecd5db7
3 years ago
Megvii Engine Team
d063d5774f
perf(functional): use fma to reduce elemwise but disable subgraph compilation
GitOrigin-RevId: c75a6e1a09
3 years ago
Megvii Engine Team
2a063f8e87
fix(subgraph): fix scope mismatch of subgraph content
GitOrigin-RevId: 6e23456250
3 years ago
Megvii Engine Team
3206af9db2
perf(functional/matmul): reimplement matmul with subgraph
GitOrigin-RevId: 456b2a51d3
3 years ago
Megvii Engine Team
8c47c1f149
perf(syncbn): reimplement with subgraph
GitOrigin-RevId: 13e7e3d3c0
3 years ago
Megvii Engine Team
53da5c79f4
feat(cg): add comp_seq_sync_device option
GitOrigin-RevId: c2199c59e9
3 years ago
Megvii Engine Team
e1c7b22ff0
perf(ops): enable memory forward for reduce in special cases
GitOrigin-RevId: dd6e1664c5
3 years ago
Megvii Engine Team
cd60d26852
perf(ops): specialize Broadcast
GitOrigin-RevId: 0cba3e6e93
3 years ago