Jake Hillion
|
ebf0e18c0e
|
re-add logos
|
2025-12-18 14:26:27 +00:00 |
|
Jake Hillion
|
2c16e00be9
|
github docs
|
2025-12-18 13:49:07 +00:00 |
|
Jake Hillion
|
f64d17fac0
|
exo v1
|
2025-12-18 13:46:40 +00:00 |
|
Jake Hillion
|
0fcee70833
|
prep repo for v1
|
2025-12-17 15:31:02 +00:00 |
|
Evan Quiney
|
09593c5e85
|
backport the dashboard to staging
|
2025-12-17 12:22:22 +00:00 |
|
Evan Quiney
|
880a18d205
|
fix disconnects
Co-authored-by: Ryuichi Leo Takashige <leo@exolabs.net>
|
2025-12-15 15:23:13 +00:00 |
|
rltakashige
|
70298ce0a9
|
Negative index nack request
|
2025-12-09 07:57:28 -08:00 |
|
Jake Hillion
|
ac3a0a6b47
|
ci: enable ruff check in CI through nix
|
2025-12-09 12:26:56 +00:00 |
|
rltakashige
|
859233a279
|
Reduce RequestEventLog spam
|
2025-12-09 11:43:54 +00:00 |
|
Evan Quiney
|
c9e2062f6e
|
switch from uvicorn to hypercorn
|
2025-12-05 17:29:06 +00:00 |
|
Jake Hillion
|
e8566a3f95
|
placement: pass different ibv_coordinator per node
|
2025-12-05 17:23:22 +00:00 |
|
Jake Hillion
|
39d76aa0a5
|
nix: move formatting checks to nix and enable in ci
|
2025-12-05 17:00:33 +00:00 |
|
Jake Hillion
|
5629983809
|
fmt: format all python/rust/nix files
|
2025-12-05 16:58:55 +00:00 |
|
Evan Quiney
|
7312a7e000
|
plan fix
|
2025-12-05 16:43:11 +00:00 |
|
Evan Quiney
|
9e0a1c23ef
|
rename ibv to jaccl inline with mlx
|
2025-12-05 16:42:43 +00:00 |
|
Evan Quiney
|
f5783d6455
|
proper collection of rdma ports in placement
|
2025-12-05 16:42:20 +00:00 |
|
Evan Quiney
|
e702313b32
|
pingers
Co-authored-by: Jake Hillion <jake@hillion.co.uk>
|
2025-12-05 16:41:19 +00:00 |
|
Evan
|
a3f8ecba9e
|
prioritise LL4
|
2025-12-05 15:08:18 +00:00 |
|
Jake Hillion
|
5ef1df1e10
|
rust: move Cargo.toml to the root
|
2025-12-05 12:01:44 +00:00 |
|
Evan
|
40a0d47de8
|
jaccl
|
2025-12-03 13:53:12 +00:00 |
|
rltakashige
|
2b243bd80e
|
Consolidate!!! Fixes
|
2025-12-03 12:19:25 +00:00 |
|
Evan Quiney
|
10c905c8dd
|
worker no longer gets stuck after shutdown
|
2025-12-02 11:35:02 +00:00 |
|
Evan
|
93f699b660
|
add aarch64-linux for the spark
|
2025-11-28 11:08:18 +00:00 |
|
Alex Cheema
|
b43d30563d
|
todo for layer-independent parameters in get_allow_patterns
|
2025-11-27 19:26:02 +00:00 |
|
Alex Cheema
|
20d73e90cd
|
fix dashboard case sensitive model id
|
2025-11-26 18:16:32 +00:00 |
|
Alex Cheema
|
e56daa7c23
|
render download progress properly
|
2025-11-26 11:48:30 +00:00 |
|
Alex Cheema
|
63c85e1298
|
get rid of spammy Finished tokenizing log
|
2025-11-25 13:02:06 +00:00 |
|
Evan
|
7088988a65
|
bump pyo3 stub-gen
|
2025-11-25 12:13:53 +00:00 |
|
rltakashige
|
7b3e3fd66c
|
Worker tests 2
|
2025-11-21 16:42:52 +00:00 |
|
rltakashige
|
de50811313
|
Worker tests on staging 1
Test plan
|
2025-11-21 15:22:40 +00:00 |
|
rltakashige
|
b45cbdeecd
|
Consolidate cleanup
|
2025-11-21 14:54:02 +00:00 |
|
rltakashige
|
28a91787e8
|
Demo
Co-authored-by: Evan <evanev7@gmail.com>
Co-authored-by: Alex Cheema <alexcheema123@gmail.com>
|
2025-11-20 20:03:51 +00:00 |
|
Alex Cheema
|
d793f5f96c
|
fix kimi eos token ids
|
2025-11-13 18:39:14 +00:00 |
|
Evan Quiney
|
b62f68474a
|
improved master error handling
Co-authored-by: Ryuichi Leo Takashige <rl.takashige@gmail.com>
|
2025-11-11 18:04:40 +00:00 |
|
Alex Cheema
|
631cb81009
|
kimi k2 thinking
|
2025-11-11 18:03:39 +00:00 |
|
Evan Quiney
|
364087b91f
|
five billion percent better shutdown handling
|
2025-11-11 17:43:53 +00:00 |
|
Evan Quiney
|
aa519b8c03
|
Worker refactor
Co-authored-by: rltakashige <rl.takashige@gmail.com>
Co-authored-by: Alex Cheema <alexcheema123@gmail.com>
|
2025-11-10 23:31:53 +00:00 |
|
Alex Cheema
|
9058b117c0
|
pipeline parallel fix
|
2025-11-08 02:19:19 +00:00 |
|
rltakashige
|
612f58c78d
|
Revert dumb merge mistake
|
2025-11-07 02:39:08 +00:00 |
|
Evan
|
6bcac37d98
|
stop benching on all pushes
|
2025-11-06 22:26:30 +00:00 |
|
rltakashige
|
ff00b165c5
|
MLX LM type stubs
|
2025-11-06 21:59:29 +00:00 |
|
Alex Cheema
|
19e90572e6
|
set max_transmit_size on gossipsub to 1MB. Fixes large message erorr
|
2025-11-06 19:18:48 +00:00 |
|
Alex Cheema
|
e60681963f
|
show ips on dashboard
|
2025-11-06 19:18:07 +00:00 |
|
rltakashige
|
0bb621b653
|
Add mlx nn stubs
|
2025-11-06 11:59:37 +00:00 |
|
Alex Cheema
|
699fd9591e
|
fix exo scripts
|
2025-11-05 21:47:08 -08:00 |
|
rltakashige
|
6bbb6344b6
|
mlx.distributed.Group type stubs
|
2025-11-06 05:26:04 +00:00 |
|
rltakashige
|
16f724e24c
|
Update staging 14
Co-authored-by: Evan <evanev7@gmail.com>
Co-authored-by: Alex Cheema <alexcheema123@gmail.com>
Co-authored-by: David Munha Canas Correia <dmunha@MacBook-David.local>
Co-authored-by: github-actions bot <github-actions@users.noreply.github.com>
|
2025-11-05 01:44:24 +00:00 |
|
Evan Quiney
|
3b409647ba
|
Squash merge merging_clusters into tensor_parallel94
|
2025-10-31 17:41:57 +00:00 |
|
Alex Cheema
|
d46c7e6a76
|
fix race condition with downloads where it cancels the download before renaming
|
2025-10-30 19:03:23 -07:00 |
|
rltakashige
|
91c635ca7a
|
Update mlx and mlx-lm packages
Co-authored-by: Evan <evanev7@gmail.com>
|
2025-10-31 01:34:43 +00:00 |
|