-
98babed4bb
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-02 11:43:09 +00:00
-
80ba4b69f1
fix: pydantic warning validator not returning self (#2474)
NanoCode012
2025-04-02 18:40:49 +07:00
-
ce07081d6c
doc updates; config fix
Dan Saunders
2025-04-01 20:35:10 +00:00
-
220c72c0bb
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-01 19:40:43 +00:00
-
0bfa180f7d
torch 2.7.0 base image for testing (#2467)
Wing Lian
2025-04-01 15:38:26 -04:00
-
3ce43b6db9
simplifying trainer mixins and adding to rl trainers
Dan Saunders
2025-04-01 17:53:12 +00:00
-
10dab5418e
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-01 16:28:56 +00:00
-
9e22c4ca6a
fix: set rl=None during inference (#2463)
NanoCode012
2025-04-01 23:25:53 +07:00
-
990b5896bc
fix: downgrade deepspeed to fix grad checkpoint oom (#2465) [skip ci]
NanoCode012
2025-04-01 23:25:05 +07:00
-
-
-
3595cb901f
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-01 16:01:36 +00:00
-
7d0eb66b54
fixing eval for SP (#2468)
Dan Saunders
2025-04-01 11:59:08 -04:00
-
5088cae726
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-01 13:41:28 +00:00
-
df119e3724
Validation for Muon optimizer with DS/FSDP (#2464)
Wing Lian
2025-04-01 09:39:12 -04:00
-
c578c8f256
Validation for Muon optimizer with DS/FSDP
muon-validation
Wing Lian
2025-04-01 09:29:54 -04:00
-
-
-
1ae67fdd05
Built site for gh-pages
Quarto GHA Workflow Runner
2025-04-01 13:22:21 +00:00
-
f4ae8816bb
Fix: remove the numerous sequential log (#2461)
NanoCode012
2025-04-01 20:20:00 +07:00
-
9b95e06cbb
Fix(doc): Minor doc changes for peft and modal (#2462) [skip ci]
NanoCode012
2025-04-01 19:48:36 +07:00
-
e0aba74dd0
Release update 20250331 (#2460) [skip ci]
Wing Lian
2025-04-01 08:47:50 -04:00
-
be8430d321
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-31 21:17:51 +00:00
-
328d598114
gemma3 packing fixes (#2449)
Wing Lian
2025-03-31 17:15:23 -04:00
-
4d36ecc724
Sequential sample packing (#2404) [skip ci]
DreamGenX
2025-03-31 21:48:20 +02:00
-
7acf93b59f
Fix(doc): Clarify doc on attention configs and missing pad_token (#2455) [skip ci]
NanoCode012
2025-04-01 02:47:28 +07:00
-
b6fc46ada8
Updates for trl 0.16.0 - mostly for GRPO (#2437) [skip ci]
Wing Lian
2025-03-31 15:47:11 -04:00
-
71afa0897d
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-31 19:20:01 +00:00
-
b35992262e
Ray train bugfix (#2458)
Dan Saunders
2025-03-31 15:17:43 -04:00
-
1defb8a955
Merge branch 'main' into destroy-pg
destroy-pg
Dan Saunders
2025-03-31 14:36:43 -04:00
-
-
-
-
70b466aa67
ray bugfix
Dan Saunders
2025-03-31 18:35:41 +00:00
-
890b28de14
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-31 16:39:07 +00:00
-
ef6eb77cc8
destroy process group on Ctrl+C / training or eval run (#2457)
Dan Saunders
2025-03-31 12:36:47 -04:00
-
32ce167404
update
Dan Saunders
2025-03-31 14:46:15 +00:00
-
1c4cc639f5
fix nccl pg destroy warning
Dan Saunders
2025-03-31 14:32:50 +00:00
-
-
-
5c57c40993
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-31 13:16:19 +00:00
-
5410195e0b
Sequence parallelism quick follow-ups; remove ModelCallback (#2450)
Dan Saunders
2025-03-31 09:13:42 -04:00
-
1a7f048c6b
add SOAP optimizer
feat/soap-optim-v2
Wing Lian
2025-03-24 03:46:59 -04:00
-
76d26366ad
upstream updates for momentum change
Wing Lian
2025-03-24 03:39:42 -04:00
-
64fe284765
add soap optimize
Wing Lian
2025-03-24 03:28:06 -04:00
-
-
-
7888a35118
chore: remove unused log
fix/xformers
NanoCode012
2025-03-31 16:20:15 +07:00
-
873385b7d5
feat: update xformers for new attention interface
NanoCode012
2025-03-31 16:15:55 +07:00
-
-
-
5cbda3b986
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-31 06:42:32 +00:00
-
cf0c79d52e
fix: minor patches for multimodal (#2441)
NanoCode012
2025-03-31 13:40:12 +07:00
-
77380cdbcc
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-29 12:32:17 +00:00
-
4ba80a0e5a
fix streaming packing test (#2454)
Wing Lian
2025-03-29 08:30:06 -04:00
-
05da8f0e9f
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-29 03:41:25 +00:00
-
c49682132b
use offline for precached stream dataset (#2453)
Wing Lian
2025-03-28 23:39:09 -04:00
-
6093306435
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-28 23:23:17 +00:00
-
e46239f8d3
bump liger to 0.5.5 (#2448)
Wing Lian
2025-03-28 19:21:03 -04:00
-
05f03b541a
hf offline decorator for tests to workaround rate limits (#2452) [skip ci]
Wing Lian
2025-03-28 19:20:46 -04:00
-
c5c01c11d8
fix dumb mistakes
mm_mc_chat
Sunny Liu
2025-03-27 13:33:52 -04:00
-
00ebf2faf9
message key checking
Sunny Liu
2025-03-27 13:29:17 -04:00
-
641e84188b
add chat conversion for multiple choice format
Sunny Liu
2025-03-27 10:51:24 -04:00
-
-
-
262ea27856
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-26 22:17:27 +00:00
-
a4e430e7c4
add override of upstream fix for multi-gpu orpo (#2440)
Wing Lian
2025-03-26 18:14:59 -04:00
-
6cdcb8ddd5
Set the pytorch_cuda_alloc_conf env in the train module (#2447)
Wing Lian
2025-03-26 18:14:43 -04:00
-
a7811ad4a0
fix(doc): document config required to run
eval_causal_lm_metrics (#2445) [skip ci]
NanoCode012
2025-03-27 05:14:29 +07:00
-
e2da821e67
chore: minor optim changes (add apollo, improve docs, remove lion-pytorch) (#2444)
NanoCode012
2025-03-27 05:14:07 +07:00
-
2c34a4634e
feat: add CCE for gemma3, cohere, and cohere2 (#2443)
NanoCode012
2025-03-27 05:13:51 +07:00
-
0fbd202764
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-23 15:11:04 +00:00
-
a9b0733f2c
Feat: Rework multimodal support (mllama, llava, pixtral, qwen2, qwen25, gemma3, mistral3) (#2435)
NanoCode012
2025-03-23 22:08:51 +07:00
-
8dc7909473
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-23 00:35:31 +00:00
-
9f00465a5c
Feat: Add support for gemma3_text and add e2e for gemma2 (#2406)
NanoCode012
2025-03-23 07:33:21 +07:00
-
571a177bc4
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-22 21:55:39 +00:00
-
86bac48d14
cleanup for failing test (#2436)
Dan Saunders
2025-03-22 17:53:29 -04:00
-
127f9229b5
Built site for gh-pages
Quarto GHA Workflow Runner
2025-03-21 17:30:33 +00:00
-
e44953d50c
installing axolotl prior to quartodoc build (#2434)
Dan Saunders
2025-03-21 13:28:13 -04:00
-
c649d569b4
simplify by installing no deps
quartodoc-fix
Dan Saunders
2025-03-21 13:27:54 -04:00
-
b88b389b17
installing axolotl prior to quartodoc build
Dan Saunders
2025-03-21 16:52:51 +00:00
-
-
-
0bffef25d0
installing axolotl prior to quartodoc build
quartodoc
Dan Saunders
2025-03-21 16:51:02 +00:00
-
23f0c51d88
Sequence parallelism (#2412)
Dan Saunders
2025-03-21 12:43:55 -04:00
-
4ac65462f0
precommit
sequence-parallelism
Dan Saunders
2025-03-21 16:43:14 +00:00
-
ce35b2a95f
precommit
Dan Saunders
2025-03-21 11:40:48 -04:00
-
ab3b36339a
fix tests
Dan Saunders
2025-03-20 12:04:22 -04:00
-
22cfa42961
small updates
Dan Saunders
2025-03-20 02:45:53 +00:00
-
0b2c2ed68c
refactors, SP mixin
Dan Saunders
2025-03-20 01:16:16 +00:00
-
2f0b4626b9
review comments, docstrings
Dan Saunders
2025-03-19 17:35:09 +00:00
-
a26985c53c
small changes
Dan Saunders
2025-03-19 17:15:30 +00:00
-
c1a58339e8
add SP doc, review comments
Dan Saunders
2025-03-18 20:04:48 +00:00
-
411df76a97
bugfix
Dan Saunders
2025-03-17 22:57:55 +00:00
-
a09d1ccbf2
removing print statement
Dan Saunders
2025-03-17 15:32:28 +00:00
-
2727d86544
non-seq2se1 collator fix
Dan Saunders
2025-03-17 13:42:49 +00:00
-
64c203cdef
sampler / dataloader refactor
Dan Saunders
2025-03-17 03:08:39 +00:00
-
7d7042f602
test fix
Dan Saunders
2025-03-17 01:21:22 +00:00
-
d187f1f8e2
using field validator instead of model validator
Dan Saunders
2025-03-17 00:28:45 +00:00
-
1cced52719
rename file, delete another
Dan Saunders
2025-03-14 15:51:37 +00:00
-
11321b17e7
removing flash-attn from requirements.txt (in setup.py extras already)
Dan Saunders
2025-03-14 09:37:24 -04:00
-
7a1a211c99
move ring flash attn to extras with flash-attn (#2414)
Wing Lian
2025-03-14 09:28:28 -04:00
-
e1a02a32b5
fix
Dan Saunders
2025-03-14 01:58:07 +00:00
-
a6ef6c7764
fix
Dan Saunders
2025-03-14 01:42:10 +00:00
-
cb3a9e99a3
gracefully handle no ring-flash-attn
Dan Saunders
2025-03-14 01:07:25 +00:00
-
3ae47ec7de
actually isolate CLI tests
Dan Saunders
2025-03-14 00:44:10 +00:00
-
e36dc763ab
isolate cli tests
Dan Saunders
2025-03-14 00:36:58 +00:00
-
03027cf6bf
pernicious Fire CLI bugfix
Dan Saunders
2025-03-14 00:18:39 +00:00
-
0ade60d455
another import scoping change
Dan Saunders
2025-03-13 23:32:07 +00:00
-
02e1a42f04
scoping down problematic import
Dan Saunders
2025-03-13 23:30:04 +00:00
-
919b88f11b
update config.qmd and rename option
Dan Saunders
2025-03-13 23:13:37 +00:00
-
345a9dd831
removing some obvious comments
Dan Saunders
2025-03-13 23:05:27 +00:00
-
4ff97bc9d4
eval dataloader and sampler changes
Dan Saunders
2025-03-13 19:24:30 +00:00
-
d0e178d52f
remove debug logs and simplify
Dan Saunders
2025-03-13 15:47:45 +00:00
-
5731cdc0cf
fixing sample packing
Dan Saunders
2025-03-12 20:44:02 +00:00
-
b7738d57c4
working multi-group SP
Dan Saunders
2025-03-12 19:33:40 +00:00
-
698e599bf7
precommit fixes
Dan Saunders
2025-03-11 14:24:48 +00:00