 Aleksandr Borzunov
|
75ed6ac49c
Fix "Too many open files" during rebalancing
|
há 2 anos atrás |
|
Alexander Borzunov
|
f64eb3a665
Update hivemind to 1.1.2, mark `model` argument as required (#81)
|
há 2 anos atrás |
|
Alexander Borzunov
|
149f433763
Rebalance swarm when necessary (#34)
|
há 2 anos atrás |
|
Alexander Borzunov
|
640bbc38a9
Make even smaller readability changes
|
há 3 anos atrás |
|
Alexander Borzunov
|
d1b012b479
Make small readability & style changes to the instructions (#77)
|
há 3 anos atrás |
 justheuristic
|
fef48d7d99
Use bitsandbytes==0.34.0, update readme (#76)
|
há 3 anos atrás |
|
justheuristic
|
8caf1145a8
Quality of life changes: update readme, simplify run_server interface (#75)
|
há 3 anos atrás |
 Artem Chumachenko
|
1046911dea
Add prompt tuning example on Personachat dataset (#69)
|
há 3 anos atrás |
|
justheuristic
|
3fdcc55a56
fix protobuf version (#74)
|
há 3 anos atrás |
|
justheuristic
|
e92487e5d2
Update dependency versions (#71)
|
há 3 anos atrás |
 Pavel Samygin
|
50535a8435
Priority tasks (#47)
|
há 3 anos atrás |
|
justheuristic
|
892d18fea7
Build cpuonly from bitsandbytes main (#70)
|
há 3 anos atrás |
|
justheuristic
|
f3984b192a
Make attention cache wait until memory is freed (#53)
|
há 3 anos atrás |
|
justheuristic
|
8a0c056929
Fix calling rpc_info multiple times (#60)
|
há 3 anos atrás |
|
Artem Chumachenko
|
ada98a1b37
Add deep prompt inference (#66)
|
há 3 anos atrás |
|
Alexander Borzunov
|
54ad745bed
Warn that current instructions involve 6B model but we will replace them soon (#63)
|
há 3 anos atrás |
|
Alexander Borzunov
|
5f0c5329d4
Update readme with arxiv link and more discussions (#62)
|
há 3 anos atrás |
|
Alexander Borzunov
|
9bea7b9ea8
Update bullet points with feedback from Tim and other people (#61)
|
há 3 anos atrás |
|
Alexander Borzunov
|
7653562aa1
Use latest version of Petals scheme, shrink Petals logo (#59)
|
há 3 anos atrás |
|
Alexander Borzunov
|
2eb5843852
Update readme for the 1st public release (#57)
|
há 3 anos atrás |
|
Pavel Samygin
|
0be21775af
remove transformer block, implement as sequential of size 1 (#54)
|
há 3 anos atrás |
|
Artem Chumachenko
|
77220c718c
Add shallow prefix-tuned inference (#55)
|
há 3 anos atrás |
|
justheuristic
|
d271b75dd4
Let users specify sequence length instead of assuming 2048 (#52)
|
há 3 anos atrás |
 Dmitry Baranchuk
|
948877149c
Fix recovering for sequential_backward (#50)
|
há 3 anos atrás |
|
Dmitry Baranchuk
|
24ba3433e4
[Fix] make distributed seq cls to not create the full bloom model (#49)
|
há 3 anos atrás |
|
justheuristic
|
f12d0deee9
[quickfix 1/n] remove expensive assertions in inference code (#48)
|
há 3 anos atrás |
|
Dmitry Baranchuk
|
0fd2caa4be
Convert actual model weights (#46)
|
há 3 anos atrás |
|
justheuristic
|
a2634001e9
Reduce vocabulary size in test model, fix bug in routing when overlapped (#45)
|
há 3 anos atrás |
|
Dmitry Baranchuk
|
5745882c67
fix rpc_forward_stream
|
há 3 anos atrás |
|
Dmitry Baranchuk
|
6095f58681
Deep distributed prompt tuning (#42)
|
há 3 anos atrás |
