 Alexander Borzunov
|
dc0072fde1
Wait for DHT storing state OFFLINE on shutdown (#486)
|
há 2 anos atrás |
|
Alexander Borzunov
|
a26559ff65
Fix `.generate(input_ids=...)` (#485)
|
há 2 anos atrás |
|
Alexander Borzunov
|
459933f846
Remove no-op process in PrioritizedTaskPool (#484)
|
há 2 anos atrás |
|
Alexander Borzunov
|
26ebbfe8f0
Support macOS (#477)
|
há 2 anos atrás |
|
Alexander Borzunov
|
75e516a8c1
Refactor readme (#482)
|
há 2 anos atrás |
 justheuristic
|
c08d09c4d3
Rewrite MemoryCache alloc_timeout logic (#434)
|
há 2 anos atrás |
|
Alexander Borzunov
|
90840dfea2
Fix requiring transformers>=4.32.0 (#480)
|
há 2 anos atrás |
|
Alexander Borzunov
|
915b357740
Require transformers>=4.32.0 (#479)
|
há 2 anos atrás |
|
Alexander Borzunov
|
18e93afc73
Don't install cpufeature on non-x86_64 machines (#478)
|
há 2 anos atrás |
|
Alexander Borzunov
|
6967904590
Bump version to 2.1.0 (#474)
|
há 2 anos atrás |
|
Alexander Borzunov
|
df8ab09ca2
Hide excess key message (#476)
|
há 2 anos atrás |
 Artem Chumachenko
|
a14ae7334d
Update peft to 0.5.0 version (#475)
|
há 2 anos atrás |
|
Alexander Borzunov
|
a9b0e9ff1a
Support loading weights from Safetensors on server (#473)
|
há 2 anos atrás |
|
justheuristic
|
4f850996bb
Change transformers version assert (#472)
|
há 2 anos atrás |
|
justheuristic
|
9250025140
Support transformers 4.32.x (#471)
|
há 2 anos atrás |
|
justheuristic
|
adda5f8c20
Temporarily require peft<0.5.0, transformers<4.32.0 (#470)
|
há 2 anos atrás |
|
Alexander Borzunov
|
de2475f31c
Make client compatible with transformers' GenerationMixin (#464)
|
há 2 anos atrás |
|
Alexander Borzunov
|
063e94b4c8
Move SequenceManagerConfig -> ClientConfig, petals.dht_utils -> petals.utils.dht (#463)
|
há 2 anos atrás |
|
Artem Chumachenko
|
568f21dc3b
Add customizable input tensors (#445)
|
há 2 anos atrás |
|
Alexander Borzunov
|
329f7d31e8
Add `blocked_servers` argument (#462)
|
há 2 anos atrás |
|
Alexander Borzunov
|
722c4dc496
Bump version to 2.0.1.post2 (#459)
|
há 2 anos atrás |
|
Alexander Borzunov
|
056f22515a
Prioritize short inference, unmerge pools for long inference (#458)
|
há 2 anos atrás |
|
justheuristic
|
55eb36ef48
Fix missing torch.cuda.synchronize for computing throughput (#456)
|
há 2 anos atrás |
|
Alexander Borzunov
|
0e7189b3ed
benchmarks: Aggregate speed among workers, set default dtype torch32 (#454)
|
há 2 anos atrás |
|
Alexander Borzunov
|
8c546d988a
Test Llama, rebalancing, throughput eval, and all CLI scripts (#452)
|
há 2 anos atrás |
|
Alexander Borzunov
|
df6fdd2d0b
Force using --new_swarm instead of empty --initial_peers (#451)
|
há 2 anos atrás |
|
Alexander Borzunov
|
2a150770a4
Prefer longer servers for fine-tuning, exclude unreachable (#448)
|
há 2 anos atrás |
|
Alexander Borzunov
|
00d48dcbe1
Override float32 in config to bfloat16 (#431)
|
há 2 anos atrás |
|
justheuristic
|
ac9b546706
[Refactor] extract block forward, backward and inference into a separate file (#435)
|
há 2 anos atrás |
|
Alexander Borzunov
|
593d980ad8
Use bitsandbytes 0.41.1 (#442)
|
há 2 anos atrás |
