Commit Graph

78 Commits (8fb9001a2c866a6796c8a34f3e7bb9d2d99ddb6a)

Author SHA1 Message Date
bmaltais 2ce5658302 Add Flux.1 Model checkbox and detection 2024-08-10 14:55:39 -04:00
bmaltais 6c8bd3c4c3 Merge branch 'dev' into sd3 2024-08-08 20:49:39 -04:00
bmaltais 630cec8587
Merge pull request #2690 from b-fission/autodetect-modeltype
Auto-detect model type for safetensors files
2024-08-08 20:26:22 -04:00
b-fission 5ccf90e7fa Auto-detect model type for safetensors files
Automatically tick the checkboxes for v2 and SDXL on the common training UI
and LoRA extract/merge utilities.
2024-08-05 19:27:24 -05:00
bmaltais 9a818cf603 Merge branch 'dev' into 173-error-on-train-typeerror-int-argument-must-be-a-string-a-bytes-like-object-or-a-real-number-not-nonetype 2024-08-02 10:54:24 -04:00
bmaltais 3121d5ec35 Fix issue with missing key. Upgrade Gradio release for security issue. 2024-08-02 10:41:13 -04:00
bmaltais 65ac2f163a Adding SD3 GUI elements 2024-07-07 15:14:53 -04:00
bmaltais 441721ab87 Use output model name as wandb run name if not provided 2024-06-24 10:26:54 -04:00
bmaltais ff61f3a4be Improve "print command" output format 2024-06-23 13:52:21 -04:00
bmaltais 018ce04852 Fix adaptive_noise_scale value not properly loading from json config 2024-06-22 08:51:12 -04:00
Lucas Freire Sangoi bfc2856c68
Minor fixes (#2480)
* Update common_gui.py

Little fix to the `validate_model_path` function to properly handle a folder path. Currently, it errors out when a folder path is selected for a training with diffusers. 😊

* Update lora_gui.py

Switching the validation type for the resume training state path from 'file' to 'folder'.
😊

* Update common_gui.py

format

* Update common_gui.py

My last fix was wrong and was returning errors when using a default model or when the VAE was not defined. This new fix works with all possibilities.
However, when a diffusers folder is used, `validate_file_path` will show a failed check, but the operation will still succeed because `validate_folder_path` will return successfully.
2024-05-10 20:08:52 -04:00
bmaltais 27b58a79f2 Fix creation of output folders 2024-05-09 13:30:27 -04:00
bmaltais 92a01a3890 Fix issue with tensorboard 2024-05-09 13:20:20 -04:00
bmaltais 7fac1e89dd Move env var config to common_gui 2024-05-07 07:58:11 -04:00
bmaltais 5984c87d22 Update model validation code 2024-05-07 07:17:27 -04:00
bmaltais 36071cc244
Improve files and folders validation (#2429) 2024-05-01 08:29:31 -04:00
bmaltais 91350e5581 Update utility code 2024-04-30 20:09:19 -04:00
bmaltais 7bbc99d91b
2382 some incorrect usage related to the recent shell=false issue (#2417)
* Convert script back to no shell.

* Update rest of tools and update dreambooth to no shell

* Update rest of trainers to not use shell

* Allow the use of custom caption extension
2024-04-29 07:44:55 -04:00
bmaltais a97d3a94ee Fix issue with pre and post fix caption in subfolders 2024-04-29 07:13:34 -04:00
bmaltais d8a51f34fd Set `max_train_steps` to 0 if not specified in older `.json` config files 2024-04-26 07:11:07 -04:00
bmaltais 1f5f3cf027 Prevent tkinter import crash 2024-04-25 11:20:27 -04:00
bmaltais 8234e52ded
Add validation of lr scheduler and optimizer arguments (#2358) 2024-04-20 15:53:11 -04:00
bmaltais 58e57a3365
Fix issue with lora_network_weights not being loaded (#2357) 2024-04-20 15:05:35 -04:00
bmaltais 5a801649c2 Convert config with use_wandb to log_with = wandb 2024-04-19 13:16:23 -04:00
bmaltais 542af98980 Fix wrong default for max_train_steps and max_train_epoch 2024-04-19 13:04:02 -04:00
bmaltais 4c5f4dfd27 update file read and write to utf-8 2024-04-19 10:50:43 -04:00
bmaltais 84fb7ce8c9 Update utility paths and print toml file when clicking the Print training command button 2024-04-17 19:54:56 -04:00
Bernard Maltais be5d30130f Fix clip_skip issue 2024-04-17 13:25:55 -04:00
bmaltais e6a8dec98d
Implement use_shell as parameter (#2297) 2024-04-15 20:26:09 -04:00
bmaltais a8320e3f2c Add support for accelerate dynamo parameters 2024-04-15 10:31:05 -04:00
bmaltais c4e0655a7a Fix accelerate path for linux systems 2024-04-15 08:50:48 -04:00
bmaltais eec6f9baf4 Convert str numbers to proper int or float 2024-04-15 08:02:39 -04:00
bmaltais a08c7c6ef1 Update parameters handling 2024-04-14 19:10:26 -04:00
bmaltais e98fc40a63
- Implement HuggingFace inputs in all training tabs (#2287)
- Switch to sd-scripts configuration files to pass most of the parameters in a secure way.
2024-04-14 18:43:45 -04:00
bmaltais fa5c65eade Update readme 2024-04-13 21:22:52 -04:00
bmaltais fddac3362e Make Stop training diseapear when training is done 2024-04-13 21:10:31 -04:00
bmaltais 4bfa486ed7 Improving args style 2024-04-13 16:46:40 -04:00
bmaltais 80afccc0e4 More fixes 2024-04-13 14:36:07 -04:00
bmaltais 1e7a263f79 Fix checkpoint name and sample prompt path 2024-04-13 14:23:46 -04:00
Bernard Maltais 835b6505c0 Fix running on linux 2024-04-13 14:03:55 -04:00
bmaltais 763eda2748 Update requirements for gradio
More shlex.quote fix
2024-04-13 12:21:26 -04:00
bmaltais 7581c4949f More fixes 2024-04-13 11:47:52 -04:00
bmaltais 62aa7ba8b6 Fix paths 2024-04-13 11:09:03 -04:00
bmaltais 79a78246a2 Rewrite run_cmd to list 2024-04-13 10:42:14 -04:00
bmaltais fef3dd2d12 Quick fix for other trainers 2024-04-13 08:43:57 -04:00
bmaltais f34eea41ca Fix issue for Dreambooth training... quick fix 2024-04-13 08:29:00 -04:00
bmaltais 25bb1303ff
Restrict extensions pre_postfix can use to improve security (#2259) 2024-04-11 08:55:03 -04:00
bmaltais 8bc67a7467
Find replace (#2258)
* Improve security to limit finding and replacing contents of any files on the system, that the attacker knows the contents of by specifying the list of acceptable extensions for captions.

* Update caption extension fields to Dropdown with restricted choices for security reasons
2024-04-11 08:32:17 -04:00
bmaltais 572dcd10ab Add Huber loss support 2024-04-07 10:35:07 -04:00
bmaltais 7b74dbfff8 Add code to check config file folder path existance 2024-04-05 07:21:17 -04:00