- 🎉🧪 feature: Graphical Web Interface [StableStudio](https://github.com/Stability-AI/StableStudio). run `aimg server` and visit http://127.0.0.1:8000/
- 🎉🧪 feature: Graphical Web Interface [StableStudio](https://github.com/Stability-AI/StableStudio). run `aimg server` and visit http://127.0.0.1:8000/
@ -103,12 +116,15 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- build: check for torch version at runtime (fixes #329)
- build: check for torch version at runtime (fixes #329)
**12.0.3**
**12.0.3**
- fix: exclude broken versions of timm as dependencies
- fix: exclude broken versions of timm as dependencies
**12.0.2**
**12.0.2**
- fix: move normal map preprocessor for conda compatibility
- fix: move normal map preprocessor for conda compatibility
**12.0.1**
**12.0.1**
- fix: use correct device for depth images on mps. Fixes #300
- fix: use correct device for depth images on mps. Fixes #300
**12.0.0**
**12.0.0**
@ -123,6 +139,7 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- fix: filenames start numbers after latest image, even if some previous images were deleted
- fix: filenames start numbers after latest image, even if some previous images were deleted
**11.1.1**
**11.1.1**
- fix: fix globbing bug with input image path handling
- fix: fix globbing bug with input image path handling
- fix: changed sample to True to generate caption using blip model
- fix: changed sample to True to generate caption using blip model
@ -135,6 +152,7 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- fix: fix model downloads that were broken by [library change in transformers 4.27.0](https://github.com/huggingface/transformers/commit/8f3b4a1d5bd97045541c43179efe8cd9c58adb76)
- fix: fix model downloads that were broken by [library change in transformers 4.27.0](https://github.com/huggingface/transformers/commit/8f3b4a1d5bd97045541c43179efe8cd9c58adb76)
**11.0.0**
**11.0.0**
- all these changes together mean same seed/sampler will not be guaranteed to produce same image (thus the version bump)
- all these changes together mean same seed/sampler will not be guaranteed to produce same image (thus the version bump)
- fix: image composition didn't work very well. Works well now but probably very slow on non-cuda platforms
- fix: image composition didn't work very well. Works well now but probably very slow on non-cuda platforms
- fix: remove upscaler tiling message
- fix: remove upscaler tiling message
@ -142,6 +160,7 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- fix: img2img was broken for all samplers except plms and ddim when init image strength was >~0.25
- fix: img2img was broken for all samplers except plms and ddim when init image strength was >~0.25
**10.2.0**
**10.2.0**
- feature: input raw control images (a pose, canny map, depth map, etc) directly using `--control-image-raw`
- feature: input raw control images (a pose, canny map, depth map, etc) directly using `--control-image-raw`
This is opposed to current behavior of extracting the control signal from an input image via `--control-image`
This is opposed to current behavior of extracting the control signal from an input image via `--control-image`
- feature: `aimg model-list` command lists included models
- feature: `aimg model-list` command lists included models
@ -153,11 +172,13 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- docs: pypi docs now link properly to github automatically
- docs: pypi docs now link properly to github automatically
**10.1.0**
**10.1.0**
- feature: 🎉 ControlNet integration! Control the structure of generated images.
- feature: 🎉 ControlNet integration! Control the structure of generated images.
- feature: `aimg colorize` attempts to use controlnet to colorize images
- feature: `aimg colorize` attempts to use controlnet to colorize images
- feature: `--caption-text` command adds text at the bottom left of an image
- feature: `--caption-text` command adds text at the bottom left of an image
**10.0.1**
**10.0.1**
- fix: `edit` was broken
- fix: `edit` was broken
**10.0.0**
**10.0.0**
@ -171,9 +192,11 @@ MacOS M1, [torch will not be able to use the M1 when generating images.](https:/
- perf: sliced latent decoding - now possible to make much bigger images. 3310x3310 on 11 GB GPU.
- perf: sliced latent decoding - now possible to make much bigger images. 3310x3310 on 11 GB GPU.
**9.0.2**
**9.0.2**
- fix: edit interface was broken
- fix: edit interface was broken
**9.0.1**
**9.0.1**
- fix: use entry_points for windows since setup.py scripts doesn't work on windows [#239](https://github.com/brycedrennan/imaginAIry/issues/239)
- fix: use entry_points for windows since setup.py scripts doesn't work on windows [#239](https://github.com/brycedrennan/imaginAIry/issues/239)
**9.0.0**
**9.0.0**
@ -187,9 +210,11 @@ batch editing of images as requested in [#229](https://github.com/brycedrennan/i
- docs: add directions on how to change model cache path
- docs: add directions on how to change model cache path
**8.3.1**
**8.3.1**
- fix: init-image-strength type
- fix: init-image-strength type
**8.3.0**
**8.3.0**
- feature: create `gifs` or `mp4s` from any images made in a single run with `--compilation-anim gif`
- feature: create `gifs` or `mp4s` from any images made in a single run with `--compilation-anim gif`
- feature: create a series of images or edits by iterating over a parameter with the `--arg-schedule` argument
- feature: create a series of images or edits by iterating over a parameter with the `--arg-schedule` argument
- feature: `openjourney-v1` and `openjourney-v2` models added. available via `--model openjourney-v2`
- feature: `openjourney-v1` and `openjourney-v2` models added. available via `--model openjourney-v2`
@ -199,9 +224,11 @@ batch editing of images as requested in [#229](https://github.com/brycedrennan/i
- fix: tile mode was broken since latest perf improvements
- fix: tile mode was broken since latest perf improvements
**8.2.0**
**8.2.0**
- feature: added `aimg system-info` command to help debug issues
- feature: added `aimg system-info` command to help debug issues
**8.1.0**
**8.1.0**
- feature: some memory optimizations and documentation
- feature: some memory optimizations and documentation
- feature: surprise-me improvements
- feature: surprise-me improvements
- feature: image sizes can now be multiples of 8 instead of 64. Inputs will be silently rounded down.
- feature: image sizes can now be multiples of 8 instead of 64. Inputs will be silently rounded down.
@ -211,44 +238,55 @@ batch editing of images as requested in [#229](https://github.com/brycedrennan/i
- fix: make captioning work with alpha pngs
- fix: make captioning work with alpha pngs
**8.0.5**
**8.0.5**
- fix: bypass huggingface cache retrieval bug
- fix: bypass huggingface cache retrieval bug
**8.0.4**
**8.0.4**
- fix: limit attention slice size on MacOS machines with 64gb (#175)
- fix: limit attention slice size on MacOS machines with 64gb (#175)
**8.0.3**
**8.0.3**
- fix: use python 3.7 compatible lru_cache
- fix: use python 3.7 compatible lru_cache
- fix: use windows compatible filenames
- fix: use windows compatible filenames
**8.0.2**
**8.0.2**
- fix: hf_hub_download() got an unexpected keyword argument 'token'
- fix: hf_hub_download() got an unexpected keyword argument 'token'
**8.0.1**
**8.0.1**
- fix: spelling mistake of "surprise"
- fix: spelling mistake of "surprise"
**8.0.0**
**8.0.0**
- feature: 🎉 edit images with instructions alone!
- feature: 🎉 edit images with instructions alone!
- feature: when editing an image add `--gif` to create a comparision gif
- feature: when editing an image add `--gif` to create a comparision gif
- feature: `aimg edit --surprise-me --gif my-image.jpg` for some fun pre-programmed edits
- feature: `aimg edit --surprise-me --gif my-image.jpg` for some fun pre-programmed edits
- feature: prune-ckpt command also removes the non-ema weights
- feature: prune-ckpt command also removes the non-ema weights
**7.6.0**
**7.6.0**
- fix: default model config was broken
- fix: default model config was broken
- feature: print version with `--version`
- feature: print version with `--version`
- feature: ability to load safetensors
- feature: ability to load safetensors
- feature: 🎉 outpainting. Examples: `--outpaint up10,down300,left50,right50` or `--outpaint all100` or `--outpaint u100,d200,l300,r400`
- feature: 🎉 outpainting. Examples: `--outpaint up10,down300,left50,right50` or `--outpaint all100` or `--outpaint u100,d200,l300,r400`
**7.4.3**
**7.4.3**
- fix: handle old pytorch lightning imports with a graceful failure (fixes #161)
- fix: handle old pytorch lightning imports with a graceful failure (fixes #161)
Previously the `+` and `-` characters in a mask (example: `face{+0.1}`) added to the grayscale value of any masked areas. This wasn't very useful. The new behavior is that the mask will expand or contract by the number of pixel specified. The technical terms for this are dilation and erosion. This allows much greater control over the masked area.
Previously the `+` and `-` characters in a mask (example: `face{+0.1}`) added to the grayscale value of any masked areas. This wasn't very useful. The new behavior is that the mask will expand or contract by the number of pixel specified. The technical terms for this are dilation and erosion. This allows much greater control over the masked area.
@ -325,39 +376,49 @@ inpainting model will automatically be used for any image-masking task
- ci: minor logging improvements
- ci: minor logging improvements
**3.0.1**
**3.0.1**
- fix: k-samplers were broken
- fix: k-samplers were broken
**3.0.0**
**3.0.0**
- feature: improved safety filter
- feature: improved safety filter
**2.4.0**
**2.4.0**
- 🎉 feature: prompt expansion
- 🎉 feature: prompt expansion
- feature: make (blip) photo captions more descriptive
- feature: make (blip) photo captions more descriptive
**2.3.1**
**2.3.1**
- fix: face fidelity default was broken
- fix: face fidelity default was broken
**2.3.0**
**2.3.0**
- feature: model weights file can be specified via `--model-weights-path` argument at the command line
- feature: model weights file can be specified via `--model-weights-path` argument at the command line
- fix: set face fidelity default back to old value
- fix: set face fidelity default back to old value
- fix: handle small images without throwing exception. credit to @NiclasEriksen
- fix: handle small images without throwing exception. credit to @NiclasEriksen
- docs: add setuptools-rust as dependency for macos
- docs: add setuptools-rust as dependency for macos
**2.2.1**
**2.2.1**
- fix: init image is fully ignored if init-image-strength = 0
- fix: init image is fully ignored if init-image-strength = 0
**2.2.0**
**2.2.0**
- feature: face enhancement fidelity is now configurable
- feature: face enhancement fidelity is now configurable
**2.1.0**
**2.1.0**
- [improved masking accuracy from clipseg](https://github.com/timojl/clipseg/issues/8#issuecomment-1259150865)
- [improved masking accuracy from clipseg](https://github.com/timojl/clipseg/issues/8#issuecomment-1259150865)
**2.0.3**
**2.0.3**
- fix memory leak in face enhancer
- fix memory leak in face enhancer
- fix blurry inpainting
- fix blurry inpainting
- fix for pillow compatibility
- fix for pillow compatibility
**2.0.0**
**2.0.0**
- 🎉 fix: inpainted areas correlate with surrounding image, even at 100% generation strength. Previously if the generation strength was high enough the generated image
- 🎉 fix: inpainted areas correlate with surrounding image, even at 100% generation strength. Previously if the generation strength was high enough the generated image
would be uncorrelated to the rest of the surrounding image. It created terrible looking images.
would be uncorrelated to the rest of the surrounding image. It created terrible looking images.
- 🎉 feature: interactive prompt added. access by running `aimg`
- 🎉 feature: interactive prompt added. access by running `aimg`
@ -370,35 +431,44 @@ would be uncorrelated to the rest of the surrounding image. It created terrible
- fix: img2img algorithm was wrong and wouldn't at values close to 0 or 1
- fix: img2img algorithm was wrong and wouldn't at values close to 0 or 1
**1.6.2**
**1.6.2**
- fix: another bfloat16 fix
- fix: another bfloat16 fix
**1.6.1**
**1.6.1**
- fix: make sure image tensors come to the CPU as float32 so there aren't compatibility issues with non-bfloat16 cpus
- fix: make sure image tensors come to the CPU as float32 so there aren't compatibility issues with non-bfloat16 cpus
**1.6.0**
**1.6.0**
- fix: *maybe* address #13 with `expected scalar type BFloat16 but found Float`
- fix: *maybe* address #13 with `expected scalar type BFloat16 but found Float`
- at minimum one can specify `--precision full` now and that will probably fix the issue
- at minimum one can specify `--precision full` now and that will probably fix the issue
- feature: tile mode can now be specified per-prompt
- feature: tile mode can now be specified per-prompt
**1.5.3**
**1.5.3**
- fix: missing config file for describe feature
- fix: missing config file for describe feature
**1.5.1**
**1.5.1**
- img2img now supported with PLMS (instead of just DDIM)
- img2img now supported with PLMS (instead of just DDIM)
- added image captioning feature `aimg describe dog.jpg` => `a brown dog sitting on grass`
- added image captioning feature `aimg describe dog.jpg` => `a brown dog sitting on grass`
- added new commandline tool `aimg` for additional image manipulation functionality
- added new commandline tool `aimg` for additional image manipulation functionality
**1.4.0**
**1.4.0**
- support multiple additive targets for masking with `|` symbol. Example: "fruit|stem|fruit stem"
- support multiple additive targets for masking with `|` symbol. Example: "fruit|stem|fruit stem"