 Alexander Borzunov
|
722c4dc496
Bump version to 2.0.1.post2 (#459)
|
il y a 2 ans |
|
Alexander Borzunov
|
056f22515a
Prioritize short inference, unmerge pools for long inference (#458)
|
il y a 2 ans |
 justheuristic
|
55eb36ef48
Fix missing torch.cuda.synchronize for computing throughput (#456)
|
il y a 2 ans |
|
Alexander Borzunov
|
0e7189b3ed
benchmarks: Aggregate speed among workers, set default dtype torch32 (#454)
|
il y a 2 ans |
|
Alexander Borzunov
|
8c546d988a
Test Llama, rebalancing, throughput eval, and all CLI scripts (#452)
|
il y a 2 ans |
|
Alexander Borzunov
|
df6fdd2d0b
Force using --new_swarm instead of empty --initial_peers (#451)
|
il y a 2 ans |
|
Alexander Borzunov
|
2a150770a4
Prefer longer servers for fine-tuning, exclude unreachable (#448)
|
il y a 2 ans |
|
Alexander Borzunov
|
00d48dcbe1
Override float32 in config to bfloat16 (#431)
|
il y a 2 ans |
|
justheuristic
|
ac9b546706
[Refactor] extract block forward, backward and inference into a separate file (#435)
|
il y a 2 ans |
|
Alexander Borzunov
|
593d980ad8
Use bitsandbytes 0.41.1 (#442)
|
il y a 2 ans |
|
Alexander Borzunov
|
32fbab5192
Remove deprecated comment in fine-tuning notebook (#443)
|
il y a 2 ans |
|
Alexander Borzunov
|
b58141ef66
Remove distracting links from readme (#441)
|
il y a 2 ans |
|
Alexander Borzunov
|
679397df0c
Update Discord links from channels to forums (#440)
|
il y a 2 ans |
 Vadim Peretokin
|
d0b5af34cd
Fix typo and make blocks message more informative (#437)
|
il y a 2 ans |
|
Alexander Borzunov
|
a1f7791d5e
Fix petals.utils.ping for servers with client-mode DHT (#430)
|
il y a 2 ans |
|
Alexander Borzunov
|
351e96bc46
Penalize servers that use relays during rebalancing (#428)
|
il y a 2 ans |
|
Alexander Borzunov
|
6a1b8a6a90
Add Stable Beluga 2 to readme (#424)
|
il y a 2 ans |
|
Alexander Borzunov
|
44fefa5e54
Add connect_timeout (#423)
|
il y a 2 ans |
|
Alexander Borzunov
|
cdc0f70653
Add Discord badge and more Discord links to readme (#422)
|
il y a 2 ans |
 Guocheng
|
8072cd9d1b
Fix stale link (#418)
|
il y a 2 ans |
|
Alexander Borzunov
|
f3fafd14a4
Bump version to 2.0.1 (#411)
|
il y a 2 ans |
|
Alexander Borzunov
|
fd19c21859
Update --update_period and --expiration defaults (#410)
|
il y a 2 ans |
|
Alexander Borzunov
|
ffb20b585c
Update commands for hosting Llama 2 in readme (#409)
|
il y a 2 ans |
|
Alexander Borzunov
|
48c6b6d963
Update README.md (#407)
|
il y a 2 ans |
|
Alexander Borzunov
|
c153cba1fa
Add Llama 2, WSL instructions to readme (#406)
|
il y a 2 ans |
|
justheuristic
|
5af04524dd
Split long sequences into chunks (#403)
|
il y a 2 ans |
|
Alexander Borzunov
|
30b94ef18b
If speedtest fails, assume network speed of 100 Mbit/s (#404)
|
il y a 2 ans |
|
Alexander Borzunov
|
8666653cf5
Fix routing through relay, default network RPS, --token, logging, readme (#399)
|
il y a 2 ans |
|
Alexander Borzunov
|
eb0664b993
Support Python 3.11 (#393)
|
il y a 2 ans |
|
Alexander Borzunov
|
6e4ebb94d2
Fix deadlocks in MemoryCache (#396)
|
il y a 2 ans |
