c65272052a
aria2 now downloads to different filename and renames afterwards
...
This is to match the behaviour of the original transformers downloader
in order to deal with the rare case of someone downloading a model using
aria2, cancelling before it finishes, and then attempting to resume the
download with the normal transformers downloader.
2022-05-11 15:45:38 -04:00
6d27084e8a
Better Aria2 Defaults
...
Trunc prevents slow allocation on windows, force_download=True has proven a more reliable default. Since models are converted to local formats it does not impact local users. And because -c is used the impact of checking if the model is correct is desirable and minimal.
2022-05-11 21:38:33 +02:00
7a3f865e3f
Prevent aria2 from resuming cancelled downloads
...
Resumed downloads tend to be very slow.
The original transformers downloader didn't allow resuming downloads
either.
2022-05-11 15:14:37 -04:00
c81f3bd084
Use `--file-allocation=trunc` instead of `--file-allocation=none`
2022-05-11 14:51:43 -04:00
f96c878d83
Use aria2 even when all model files are already in cache
...
This allows aria2 to continue downloading a pytorch_model.bin after a
cancelled download.
2022-05-11 14:43:56 -04:00
f60c7d8492
Fix the behaviour of `aria2_hook()` when using `force_download`
2022-05-11 14:41:34 -04:00
5732a8f15a
Don't use `aria2_hook()` if `force_download=True` is used
2022-05-11 14:40:31 -04:00
903d593ce4
Merge pull request #125 from VE-FORBRYDERNE/aria2
...
Use aria2 to improve HF model download speeds in Colab
2022-05-11 07:55:53 +02:00
46cfa1367f
Add `--no_aria2` command line flag
2022-05-11 00:44:56 -04:00
f09959f9be
Fix patching code of `PreTrainedModel.from_pretrained()`
2022-05-11 00:41:53 -04:00
22b4f3c9df
Bug fixes for `aria2_hook()` when running Windows
2022-05-11 00:14:00 -04:00
82205722af
Fix logic of `aria2_hook()`
2022-05-10 23:46:29 -04:00
4b49d1c464
Make sure `vars.revision` is defined
2022-05-10 22:51:36 -04:00
4b693b4858
Fix the logic of `force_download` in utils.py
2022-05-10 22:47:03 -04:00
c1ef20bcff
Also enable aria2 downloading for non-sharded checkpoints
2022-05-10 22:43:41 -04:00
e115bb68e4
aria2 downloads in utils.py now use correct user agent
2022-05-10 22:22:46 -04:00
b97b2a02d6
Add `--revision` command line flag
2022-05-10 22:14:56 -04:00
937d9ee06a
Change default `model.save_pretrained` shard size to 500 MiB
2022-05-10 22:04:25 -04:00
a388c63023
Use aria2 to download split checkpoints
2022-05-10 21:28:13 -04:00
01e15d03d6
Remove play.ipnyb
...
Interactive Python doesn't work well on Jupyter, until they support what Colab can do this file is pointless.
2022-05-11 01:49:07 +02:00
7a9297adc3
Jupyter Git integration
2022-05-11 01:31:12 +02:00
f917d3438f
Updated models
2022-05-10 21:39:16 +02:00
7fcc1a9acb
Fix C1
2022-05-10 18:38:50 +02:00
c5462ec480
Better Jupyter
2022-05-09 02:41:00 +02:00
e09b939f04
Force Bash
2022-05-08 16:02:16 +02:00
0ca4917056
Linux Runtime Info
2022-05-08 00:24:14 +02:00
030df1a09f
Small installer fix
2022-05-07 20:22:33 +02:00
a3dc188c8f
Linux Installer Improvements
2022-05-01 15:58:37 +02:00
9f7c9c4b9e
Merge pull request #124 from Crafteko/united
...
Replaced Adventure 125M and added C1-1.3B to the menu
2022-04-29 17:38:55 +02:00
a61ba0d000
Upload map file for GPT-NeoX
2022-04-29 00:41:56 -04:00
9c83ef7fa9
Replaced Adventure 125M and added C1-1.3B
2022-04-28 22:35:04 +00:00
810f6614af
Cap GIT version for now
2022-04-27 18:20:43 +02:00
716951f059
Merge pull request #123 from VE-FORBRYDERNE/settings
...
Prevent the settings throttle from lagging the sliders
2022-04-27 03:17:53 +02:00
b1faca3686
Prevent the settings throttle from lagging the sliders
2022-04-26 21:14:44 -04:00
9cff8268b5
Lower slider latency for now
2022-04-27 00:53:55 +02:00
b2277f242b
Merge pull request #122 from VE-FORBRYDERNE/settings
2022-04-27 00:49:01 +02:00
02e4e6be1e
Fix author's note slider
2022-04-26 17:29:18 -04:00
ee8ced2f5f
Allow users to type in the values for the settings
2022-04-26 15:27:28 -04:00
ea82867e4d
Merge branch 'united' into settings
2022-04-26 13:58:01 -04:00
4ce9a5fe28
New text
2022-04-26 19:45:37 +02:00
cd1a02e705
New TPU Colab
2022-04-26 19:14:40 +02:00
e94b97790c
Merge pull request #121 from VE-FORBRYDERNE/code
...
Fix the vscode notebook
2022-04-21 20:01:52 +02:00
c873d36374
Merge pull request #120 from VE-FORBRYDERNE/lazy-loader
...
Fix some lazy loader edge cases
2022-04-21 19:59:56 +02:00
e4cf19e707
Fix the vscode notebook
2022-04-21 13:13:46 -04:00
2d38e90509
Remove lm_head.weight from maps/xglm.json
2022-04-20 12:56:57 -04:00
6803531384
Force grad to be off by default when loading with lazy loader
2022-04-19 12:26:02 -04:00
a82a165146
ColabKobold Fixes
2022-04-19 15:15:57 +02:00
11280a6e66
LocalTunnel Linux Fix
2022-04-19 14:41:21 +02:00
b8e79afe5e
LocalTunnel support
2022-04-19 13:47:44 +02:00
c7b03398f6
Merge 'nolialsea/patch-1' into settings without Colab changes
2022-04-17 12:15:36 -04:00
Gnome Ann
Henk
subtlewave