 Max Ryabinin
|
5233b6c085
Split hivemind.client into hivemind.averaging and hivemind.moe (#304)
|
4 anos atrás |
 justheuristic
|
200fbecdbf
Refactor MPFuture to use a single pipe/thread per process (#298)
|
4 anos atrás |
|
justheuristic
|
0a0e290ea3
Add per-tensor compression, make all-reduce faster and more flexible (#272)
|
4 anos atrás |
 Vsevolod-pl
|
8673071bc0
Fixed nan when compressing the tensor of zeros (#266)
|
4 anos atrás |
 mponty
|
0080028e25
Add uniform compression (#202)
|
4 anos atrás |
|
Max Ryabinin
|
916c3db52d
Move compression-related code to hivemind.utils.compression (#213)
|
4 anos atrás |
|
Max Ryabinin
|
6f8f192150
Improve Runtime exception handling (#207)
|
4 anos atrás |
|
Vsevolod-pl
|
b9c02ac191
Add quantile compression (#182)
|
4 anos atrás |
 Alexey Bukhtiyarov
|
bb7c4d5953
Fix scalar deserialization (#190)
|
4 anos atrás |
|
Alexey Bukhtiyarov
|
06162992fa
Serialize DHTID source with msgpack (#172)
|
4 anos atrás |
|
justheuristic
|
23e655640c
Handle edge cases in DecentralizedAverager (#171)
|
4 anos atrás |
 Anton Sinitsin
|
0d7818b2cd
fix problem with NamedTuple inheritance in python3.9 (#142)
|
4 anos atrás |
|
justheuristic
|
e159605143
Address averaging corner cases, add benchmark_averaging.py, chunk averaged tensors, fix DHTNode get (#134)
|
4 anos atrás |
|
justheuristic
|
1754792aad
Process-wide channel cache for gRPC+aio (#120)
|
4 anos atrás |
|
justheuristic
|
2bd481c73f
added torch1.7 support, switch to grpc 1.33, grpc bump, improved tests & logging, (#116)
|
4 anos atrás |
|
Vsevolod-pl
|
fde83bba02
Implemented float16 compression (#106)
|
4 anos atrás |
|
Vsevolod-pl
|
d4d9da9d3e
Tensor compression (part1) (#102)
|
4 anos atrás |
|
justheuristic
|
535318e249
Simplify & explain hivemind.dht.DHT (#78)
|
5 anos atrás |
