justheuristic
|
9250025140
Support transformers 4.32.x (#471)
|
2 jaren geleden |
justheuristic
|
adda5f8c20
Temporarily require peft<0.5.0, transformers<4.32.0 (#470)
|
2 jaren geleden |
justheuristic
|
1e5df2916e
Merge branch 'main' into forward_kwargs
|
2 jaren geleden |
Your Name
|
d51c08ef20
undo debug change
|
2 jaren geleden |
Your Name
|
4529471f3f
black, isort
|
2 jaren geleden |
Your Name
|
13c13d347a
wip (again)
|
2 jaren geleden |
Your Name
|
65e87395bc
wip (again)
|
2 jaren geleden |
Alexander Borzunov
|
de2475f31c
Make client compatible with transformers' GenerationMixin (#464)
|
2 jaren geleden |
Your Name
|
084d565845
priority pool
|
2 jaren geleden |
Your Name
|
355c1509e1
black-isort
|
2 jaren geleden |
Your Name
|
fb9b21132c
black-isort
|
2 jaren geleden |
Your Name
|
ed8d7f41b8
mwp
|
2 jaren geleden |
Your Name
|
f313730767
WIP
|
2 jaren geleden |
Your Name
|
1879788705
typos
|
2 jaren geleden |
Alexander Borzunov
|
063e94b4c8
Move SequenceManagerConfig -> ClientConfig, petals.dht_utils -> petals.utils.dht (#463)
|
2 jaren geleden |
Artem Chumachenko
|
568f21dc3b
Add customizable input tensors (#445)
|
2 jaren geleden |
Alexander Borzunov
|
329f7d31e8
Add `blocked_servers` argument (#462)
|
2 jaren geleden |
Alexander Borzunov
|
722c4dc496
Bump version to 2.0.1.post2 (#459)
|
2 jaren geleden |
Alexander Borzunov
|
056f22515a
Prioritize short inference, unmerge pools for long inference (#458)
|
2 jaren geleden |
justheuristic
|
55eb36ef48
Fix missing torch.cuda.synchronize for computing throughput (#456)
|
2 jaren geleden |
Alexander Borzunov
|
0e7189b3ed
benchmarks: Aggregate speed among workers, set default dtype torch32 (#454)
|
2 jaren geleden |
Alexander Borzunov
|
8c546d988a
Test Llama, rebalancing, throughput eval, and all CLI scripts (#452)
|
2 jaren geleden |
Alexander Borzunov
|
df6fdd2d0b
Force using --new_swarm instead of empty --initial_peers (#451)
|
2 jaren geleden |
Alexander Borzunov
|
2a150770a4
Prefer longer servers for fine-tuning, exclude unreachable (#448)
|
2 jaren geleden |
Alexander Borzunov
|
00d48dcbe1
Override float32 in config to bfloat16 (#431)
|
2 jaren geleden |
justheuristic
|
ac9b546706
[Refactor] extract block forward, backward and inference into a separate file (#435)
|
2 jaren geleden |
Alexander Borzunov
|
593d980ad8
Use bitsandbytes 0.41.1 (#442)
|
2 jaren geleden |
Alexander Borzunov
|
32fbab5192
Remove deprecated comment in fine-tuning notebook (#443)
|
2 jaren geleden |
Alexander Borzunov
|
b58141ef66
Remove distracting links from readme (#441)
|
2 jaren geleden |
Alexander Borzunov
|
679397df0c
Update Discord links from channels to forums (#440)
|
2 jaren geleden |