Commit message (Expand) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Add contextmanager to EMAModel to apply weights temporarily | Volpeon | 2023-01-06 | 2 | -33/+36 |
* | Log EMA decay | Volpeon | 2023-01-05 | 1 | -1/+2 |
* | Added EMA to TI | Volpeon | 2023-01-05 | 4 | -10/+157 |
* | Fix LR finder | Volpeon | 2023-01-05 | 1 | -7/+23 |
* | Update | Volpeon | 2023-01-05 | 1 | -1/+1 |
* | Update | Volpeon | 2023-01-05 | 7 | -51/+75 |
* | Fix | Volpeon | 2023-01-05 | 2 | -10/+13 |
* | Various cleanups | Volpeon | 2023-01-05 | 8 | -1091/+133 |
* | Update | Volpeon | 2023-01-04 | 6 | -9/+15 |
* | Various updates | Volpeon | 2023-01-04 | 5 | -34/+87 |
* | Better eval generator | Volpeon | 2023-01-04 | 3 | -13/+13 |
* | Fixed reproducibility, more consistant validation | Volpeon | 2023-01-04 | 5 | -43/+113 |
* | Don't use vector_dropout by default | Volpeon | 2023-01-03 | 2 | -2/+2 |
* | Added vector dropout | Volpeon | 2023-01-03 | 4 | -15/+69 |
* | Fixed LR finder | Volpeon | 2023-01-02 | 3 | -22/+26 |
* | Update | Volpeon | 2023-01-02 | 3 | -9/+55 |
* | Fix | Volpeon | 2023-01-02 | 2 | -8/+2 |
* | Save args before training, too | Volpeon | 2023-01-02 | 2 | -0/+8 |
* | Fix | Volpeon | 2023-01-02 | 2 | -4/+2 |
* | Update | Volpeon | 2023-01-02 | 2 | -13/+12 |
* | Fix | Volpeon | 2023-01-02 | 1 | -1/+1 |
* | Improved one cycle scheduler | Volpeon | 2023-01-02 | 2 | -31/+63 |
* | Update | Volpeon | 2023-01-01 | 3 | -15/+15 |
* | Cleanup | Volpeon | 2023-01-01 | 3 | -14/+5 |
* | Fix MultiCLIPTokenizer (forgot to override encode) | Volpeon | 2023-01-01 | 1 | -13/+20 |
* | Updates | Volpeon | 2023-01-01 | 6 | -147/+227 |
* | Fixed accuracy calc, other improvements | Volpeon | 2023-01-01 | 5 | -60/+74 |
* | Fix | Volpeon | 2023-01-01 | 1 | -1/+1 |
* | Better token shuffling | Volpeon | 2023-01-01 | 1 | -2/+3 |
* | Fix | Volpeon | 2022-12-31 | 1 | -5/+7 |
* | Update | Volpeon | 2022-12-31 | 4 | -26/+35 |
* | Bugfixes for multi-vector token handling | Volpeon | 2022-12-31 | 4 | -27/+53 |
* | Simplified multi-vector embedding code | Volpeon | 2022-12-31 | 3 | -17/+14 |
* | Fixes | Volpeon | 2022-12-31 | 2 | -3/+2 |
* | Added multi-vector embeddings | Volpeon | 2022-12-31 | 8 | -81/+299 |
* | Misc improvements | Volpeon | 2022-12-30 | 3 | -78/+56 |
* | Training script improvements | Volpeon | 2022-12-30 | 5 | -25/+89 |
* | Update | Volpeon | 2022-12-29 | 2 | -11/+16 |
* | Training improvements | Volpeon | 2022-12-29 | 3 | -29/+42 |
* | Updated 1-cycle scheduler | Volpeon | 2022-12-28 | 2 | -9/+15 |
* | Integrated updates from diffusers | Volpeon | 2022-12-28 | 5 | -39/+59 |
* | Improved learning rate finder | Volpeon | 2022-12-27 | 3 | -23/+30 |
* | Added validation phase to learn rate finder | Volpeon | 2022-12-27 | 2 | -15/+31 |
* | Added learning rate finder | Volpeon | 2022-12-27 | 4 | -162/+257 |
* | Set default dimensions to 768; add config inheritance | Volpeon | 2022-12-26 | 7 | -26/+36 |
* | Code simplifications, avoid autocast | Volpeon | 2022-12-25 | 4 | -82/+92 |
* | Update | Volpeon | 2022-12-25 | 3 | -3/+3 |
* | Update | Volpeon | 2022-12-24 | 1 | -3/+3 |
* | Fixed Textual Inversion | Volpeon | 2022-12-24 | 3 | -13/+21 |
* | Training update | Volpeon | 2022-12-24 | 3 | -24/+8 |