hiyouga
e66b8ade4d
support unsloth generate
...
Former-commit-id: 0ef1ad9f505dba71db9342f524cc3a7565e5e09e
2024-04-24 04:46:53 +08:00
hiyouga
460da206f6
refactor patcher
...
Former-commit-id: 263cfe1294f5c3188f5e8d65791f35ee0d87315a
2024-04-24 03:02:23 +08:00
hiyouga
33f525455e
reenable sdpa and fast tok by default
...
Former-commit-id: 9e00902dbedc71d55743d1bf237843506a557891
2024-04-24 02:18:44 +08:00
hiyouga
83b8bc8937
fix #3347 #3387
...
Former-commit-id: c253c18185a29b59190f3e0ed236c2bb4c788085
2024-04-24 01:30:16 +08:00
hiyouga
04da91e84e
fix #3365
...
Former-commit-id: 415ce41e8fa887e980e5bd575c8e95bd4076b90b
2024-04-21 19:20:18 +08:00
hiyouga
366c0eb1c5
fix mod stuff
...
Former-commit-id: cf3988226e6398c67bb2955578e436fc505aa5c5
2024-04-21 18:11:10 +08:00
hoshi-hiyouga
5c3922713a
Merge pull request #3338 from astramind-ai/main
...
Adding Mixture of Depth
Former-commit-id: 4da2ece53353b63e672ff529d6beba41ff710c14
2024-04-21 18:05:52 +08:00
hoshi-hiyouga
7279a7014c
fix #3348
...
Former-commit-id: aa5e921c00f60074eceb2f9d4d8837cc713edba6
2024-04-20 10:34:09 +08:00
Marco
68dbd5d220
Added Mixture of Depths
...
Former-commit-id: 75dd98b9abc847e22cb263c17ebcd2ca5dd98345
2024-04-18 20:31:24 +02:00
hiyouga
0d0c6612a5
fix #3324
...
Former-commit-id: 5e710c4ac331f3400534d33b2646c4108c898d98
2024-04-18 15:34:45 +08:00
hiyouga
dd992dcce9
fix #3316
...
Former-commit-id: 7395e9e90a209228ff563ab54319955608850fc3
2024-04-17 22:54:34 +08:00
hoshi-hiyouga
e8667f9c90
Merge pull request #3287 from Ledzy/badam
...
[Feature] Add BAdam algorithm
Former-commit-id: 10a5e1e65b34b03e5ca2a41bf6ded09a3fb25f0c
2024-04-16 17:32:16 +08:00
hoshi-hiyouga
b5c3d23a22
Update utils.py
...
Former-commit-id: 7edf4dbed88b8034282f14fd6e0cb6f7f9e5f805
2024-04-16 17:29:30 +08:00
hoshi-hiyouga
88a080ced4
Update patcher.py
...
Former-commit-id: 494e6a1e05b38f5ff61d83327303614f53c92e64
2024-04-16 17:29:19 +08:00
hoshi-hiyouga
12f43694be
Update adapter.py
...
Former-commit-id: 8f7b75b26f020d8ae85baab7b082475c3bfeb512
2024-04-16 17:28:12 +08:00
Jonery
2ba03e6ef3
resolve gradient checkpointing issue.
...
Former-commit-id: 6df9135d063bb6102f0cbcdf0d702076f5febbae
2024-04-16 12:05:27 +08:00
hiyouga
ba4efe3ff6
support unsloth 2024.4
...
Former-commit-id: 14a83f8bc4fe44783252378fce59198194a96bb8
2024-04-16 00:25:03 +08:00
hiyouga
2aa1d1476e
add codegemma
...
Former-commit-id: 9324176525c2eda22962b0ca1895009b6237e6e3
2024-04-16 00:11:15 +08:00
hiyouga
19874e39ee
support cohere commandR #3184
...
Former-commit-id: e077c36872740f6b2ac255aee9da6c4c70f28977
2024-04-15 23:26:42 +08:00
Jonery
22188f1fa3
Feature BAdam
...
Former-commit-id: d8d2807fbcf587c37f7fd34a23e9397d2775ceed
2024-04-15 23:15:27 +08:00
hiyouga
be206df674
update examples
...
Former-commit-id: 369294b31c8a03a1cafcee83eb31a817007d3c49
2024-04-15 22:14:34 +08:00
hoshi-hiyouga
740d89e9df
Merge pull request #3276 from liu-zichen/fix_mixtral
...
fix: turn on output_router_logits of mixtral
Former-commit-id: 07bbaf5c67d00a152e5304e81b15fd9189e7bb99
2024-04-15 15:38:16 +08:00
hiyouga
506276c9cb
fix #3273
...
Former-commit-id: 3b20c89b342a068356ffc29c3724b645775c65db
2024-04-15 15:32:58 +08:00
liuzc
44c86150c9
fix: mixtral output_router_logits
...
Former-commit-id: ab3171ea97ec968b972287287ef9ee2502c6d37c
2024-04-15 12:11:49 +08:00
hiyouga
a97f8d1fa8
release v0.6.2
...
Former-commit-id: f92ad0a62d957b595f6a76a5403216b163eb3d17
2024-04-11 20:08:51 +08:00
hoshi-hiyouga
db51e05205
Update adapter.py
...
Former-commit-id: 720fde3683529ed7e08ac27c7c4598c6bdc30d44
2024-04-10 00:57:51 +08:00
hoshi-hiyouga
bfb090ed7a
Update adapter.py
...
Former-commit-id: a84b8d17dbf221259212e81931d80bcdd6284ad7
2024-04-10 00:57:30 +08:00
Erich Schubert
cc2ff3065f
Pass additional_target to unsloth
...
Fixes #3200
Former-commit-id: f8f87f5b0549cba6a011749c42064047f82ba577
2024-04-09 17:53:40 +02:00
hiyouga
f8609236ab
fix quant infer and qwen2moe
...
Former-commit-id: b75d16767f35c36e2cf2aaab8a3844135085bccf
2024-04-09 17:12:59 +08:00
hiyouga
d97150c571
fix resize vocab at inference #3022
...
Former-commit-id: c243720b89eec0af2872fa3c7980a0026d893f4d
2024-04-03 18:14:24 +08:00
hiyouga
75819c1220
simplify readme
...
Former-commit-id: 0da6ec2d516326fe9c7583ba71cd1778eb838178
2024-04-02 20:07:43 +08:00
hiyouga
76ba7b51c1
add moe aux loss control #3085
...
Former-commit-id: c9187ebc944e2de454ace3304b7d28eabb1b1a81
2024-04-02 14:26:31 +08:00
hiyouga
e645b9f42b
fix #3022
...
Former-commit-id: dac2f617bda9470ac8d85c7e9def09cc04970506
2024-04-02 13:58:39 +08:00
hiyouga
357a32d7a0
fix #3083
...
Former-commit-id: ff9a3f73961a362d0ddc22079f80a85465fffda8
2024-04-01 22:53:52 +08:00
hiyouga
8365522ce2
fix #3077
...
Former-commit-id: d0340391e8075cff0d84b3ef879c2101b66ca1dc
2024-04-01 21:35:18 +08:00
hiyouga
4e3ee3b703
support infer 4bit model on GPUs #3023
...
Former-commit-id: 950a9dab9055839990656b2b40956792b253573d
2024-04-01 17:34:04 +08:00
hiyouga
0c96919aa5
tiny fix
...
Former-commit-id: ba4a9b3c01e2f7467fbc5be268f47c0d003caa65
2024-03-31 00:10:29 +08:00
marko1616
e9060f37e4
fix blank line contains whitespace
...
Former-commit-id: 7bc3bcc64353d5a1d4870c6a9509b64cff710492
2024-03-30 23:46:55 +08:00
marko1616
fb6e653443
Fix Llama model save for full param train
...
Former-commit-id: ca17b5db4f97c3ec9fe2004877f150e8f51ab4b5
2024-03-30 23:45:04 +08:00
hiyouga
27f5c967e4
update trainers
...
Former-commit-id: d0dd6eefed0b86895ed00a7cafb331e5193db645
2024-03-28 18:16:27 +08:00
hiyouga
52eb06e2ee
fix #2961
...
Former-commit-id: 616917bb3be7f71073b56ad8c7bc4e164b08b9b5
2024-03-26 17:26:14 +08:00
hiyouga
6d7c325f19
fix #2928
...
Former-commit-id: 9558ee87bc7260a6596385aaa375df544862bfa9
2024-03-24 00:34:54 +08:00
hiyouga
06019b7ee3
fix #2941
...
Former-commit-id: 3775ab52017f0b610ddd8199cccfb8c001eda507
2024-03-24 00:28:44 +08:00
hiyouga
b590e82d41
support fsdp + qlora
...
Former-commit-id: b894bf8e84be689db258021f0638e9ac939abcbc
2024-03-21 00:36:06 +08:00
hiyouga
6302cd94c8
fix #2346
...
Former-commit-id: c8888c499b0ac51e2fc86c16e8e91c79400a5993
2024-03-20 17:56:33 +08:00
hiyouga
2e81c03f41
fix patcher
...
Former-commit-id: 6a5ad99c8cbf6b7def0a130306d49e7d1eb4e5a5
2024-03-15 19:18:42 +08:00
S3Studio
bada9f71a7
Use official Nvidia base image
...
Note that the flash-attn library is installed in this image and the qwen model will use it automatically.
However, if the the host machine's GPU is not compatible with the library, an exception will be raised during the training process as follows:
FlashAttention only supports Ampere GPUs or newer.
So if the --flash_attn flag is not set, an additional patch for the qwen model's config is necessary to set the default value of use_flash_attn from "auto" to False.
Former-commit-id: cd2f5717d676e1a5afd2f4e7a38402d2e55e7479
2024-03-15 08:59:13 +08:00
hiyouga
0d245f67ea
fix export
...
Former-commit-id: c9b968b84c97c9a00fbb43194c3adc9354d74f3b
2024-03-14 18:17:01 +08:00
hiyouga
4ef67ed4dd
improve lora+ impl.
...
Former-commit-id: 332bad25455a70ad9204e7dd384bb086d789aa39
2024-03-13 23:32:51 +08:00
hiyouga
d233c4a71b
fix #2802
...
Former-commit-id: f4c56ccd785790c02f0d1275cd75958677a18690
2024-03-13 12:33:45 +08:00
