From 6e0a3bc25d71506c0fa0cacdef0968b7f9bf402a Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Tue, 3 Sep 2024 09:50:51 +0300 Subject: [PATCH 01/23] little change to task description --- packages/tasks/src/tasks/text-to-image/data.ts | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/data.ts b/packages/tasks/src/tasks/text-to-image/data.ts index 7656c6edaf..d7a73cadba 100644 --- a/packages/tasks/src/tasks/text-to-image/data.ts +++ b/packages/tasks/src/tasks/text-to-image/data.ts @@ -92,7 +92,7 @@ const taskData: TaskDataCustom = { }, ], summary: - "Generates images from input text. These models can be used to generate and modify images based on text prompts.", + "Text-to-image is the task of generating images from input text. These pipelines can also be used modify and edit images based on text prompts.", widgetModels: ["CompVis/stable-diffusion-v1-4"], youtubeId: "", }; From 0858b9668eda2d5de2ec37204aa5d53eb759e2e1 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Tue, 3 Sep 2024 10:25:36 +0300 Subject: [PATCH 02/23] add initial task variants --- packages/tasks/src/tasks/text-to-image/about.md | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 79b00c005d..5c9ea5ff1a 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -16,9 +16,19 @@ Different patterns can be generated to obtain unique pieces of fashion. Text-to- Architects can utilise the models to construct an environment based out on the requirements of the floor plan. This can also include the furniture that has to be placed in that environment. -## Task Variants +##Task Variants -You can contribute variants of this task [here](https://github.com/huggingface/hub-docs/blob/main/tasks/src/text-to-image/about.md). +### Image Editing + +Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. +- synthetic image editing +- real image editing + +### Controlled Generation + +### Personalization +- subject driven generation +- style transfer ## Inference From 017ff04b2f40028b6606a3803836c0d5193bc5ff Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Wed, 4 Sep 2024 15:43:55 +0300 Subject: [PATCH 03/23] add image editing task --- packages/tasks/src/tasks/text-to-image/about.md | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 5c9ea5ff1a..fc85205177 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -2,7 +2,7 @@ ### Data Generation -Businesses can generate data for their their use cases by inputting text and getting image outputs. +Businesses can generate data for their use cases by inputting text and getting image outputs. ### Immersive Conversational Chatbots @@ -21,8 +21,10 @@ Architects can utilise the models to construct an environment based out on the r ### Image Editing Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. -- synthetic image editing -- real image editing +- synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. +![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) +- real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. +![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) ### Controlled Generation From baa2c37b3101afaa378aa26e14811596cb1113a1 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 14:54:13 +0300 Subject: [PATCH 04/23] add personalization --- packages/tasks/src/tasks/text-to-image/about.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index fc85205177..27fb153392 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -26,11 +26,11 @@ Image editing with text-to-image models involves using text prompts to describe - real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) -### Controlled Generation - ### Personalization -- subject driven generation -- style transfer +Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects. +For example, one can use these techniques to generate images of themselves, using as little as one reference image. +These include teaching the model a new concept both in training free manner or through fine-tuning. + ## Inference From d9ef4656f89b2b46c70e3515afdae884ff62259f Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 15:08:33 +0300 Subject: [PATCH 05/23] add links --- packages/tasks/src/tasks/text-to-image/about.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 27fb153392..a09c73171f 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -27,7 +27,7 @@ Image editing with text-to-image models involves using text prompts to describe ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) ### Personalization -Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects. +Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. @@ -77,11 +77,13 @@ await inference.textToImage({ - [Introducing Würstchen: Fast Diffusion for Image Generation](https://huggingface.co/blog/wuerstchen) - [Efficient Controllable Generation for SDXL with T2I-Adapters](https://huggingface.co/blog/t2i-sdxl-adapters) - [Welcome aMUSEd: Efficient Text-to-Image Generation](https://huggingface.co/blog/amused) - +- Image Editing Demos: [LEDITS++](https://huggingface.co/spaces/editing-images/leditsplusplus), [Turbo Edit](https://huggingface.co/spaces/turboedit/turbo_edit), [InstructPix2Pix](https://huggingface.co/spaces/timbrooks/instruct-pix2pix), [CosXL](https://huggingface.co/spaces/multimodalart/cosxl) +- Training free Personalization Demos: [Face-to-All](https://huggingface.co/spaces/multimodalart/face-to-all), [InstantStyle](https://huggingface.co/spaces/InstantX/InstantStyle), [RB-modulation](https://huggingface.co/spaces/fffiloni/RB-Modulation), [Photomaker v2](https://huggingface.co/spaces/TencentARC/PhotoMaker-V2) ### Model Fine-tuning - [Finetune Stable Diffusion Models with DDPO via TRL](https://huggingface.co/blog/pref-tuning) - [LoRA training scripts of the world, unite!](https://huggingface.co/blog/sdxl_lora_advanced_script) - [Using LoRA for Efficient Stable Diffusion Fine-Tuning](https://huggingface.co/blog/lora) +- LoRA fine tuning Spaces: [FLUX.1 finetuning](https://huggingface.co/spaces/autotrain-projects/train-flux-lora-ease), [SDXL finetuning](https://huggingface.co/spaces/multimodalart/lora-ease) This page was made possible thanks to the efforts of [Ishan Dutta](https://huggingface.co/ishandutta), [Enrique Elias Ubaldo](https://huggingface.co/herrius) and [Oğuz Akif](https://huggingface.co/oguzakif). From 47801f289a323f122d7d6601fc739bcfeeef2b19 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 15:16:55 +0300 Subject: [PATCH 06/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 1 + 1 file changed, 1 insertion(+) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index a09c73171f..d53e852262 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -21,6 +21,7 @@ Architects can utilise the models to construct an environment based out on the r ### Image Editing Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. + - synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) - real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. From b5f8250b958748ace5f735f537efe5cf74c15de6 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 15:19:11 +0300 Subject: [PATCH 07/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 1 + 1 file changed, 1 insertion(+) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index d53e852262..bb8fdaceb2 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -80,6 +80,7 @@ await inference.textToImage({ - [Welcome aMUSEd: Efficient Text-to-Image Generation](https://huggingface.co/blog/amused) - Image Editing Demos: [LEDITS++](https://huggingface.co/spaces/editing-images/leditsplusplus), [Turbo Edit](https://huggingface.co/spaces/turboedit/turbo_edit), [InstructPix2Pix](https://huggingface.co/spaces/timbrooks/instruct-pix2pix), [CosXL](https://huggingface.co/spaces/multimodalart/cosxl) - Training free Personalization Demos: [Face-to-All](https://huggingface.co/spaces/multimodalart/face-to-all), [InstantStyle](https://huggingface.co/spaces/InstantX/InstantStyle), [RB-modulation](https://huggingface.co/spaces/fffiloni/RB-Modulation), [Photomaker v2](https://huggingface.co/spaces/TencentARC/PhotoMaker-V2) + ### Model Fine-tuning - [Finetune Stable Diffusion Models with DDPO via TRL](https://huggingface.co/blog/pref-tuning) From 2c8b759c3ec77651f74c2ec7ca14371d0329b332 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 15:26:01 +0300 Subject: [PATCH 08/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index bb8fdaceb2..ecb5b2b771 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -23,11 +23,15 @@ Architects can utilise the models to construct an environment based out on the r Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. - synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. -![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) + +- ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) + - real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. -![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) + +- ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) ### Personalization + Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. From 812f8e2ecc05dbe1f5b8d405f50fd49bae998848 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Thu, 5 Sep 2024 15:29:04 +0300 Subject: [PATCH 09/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index ecb5b2b771..3078eda0d0 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -20,22 +20,19 @@ Architects can utilise the models to construct an environment based out on the r ### Image Editing -Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. +Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. -- synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. +- Synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. -- ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) + ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) -- real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. +- Real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. -- ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) + ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) ### Personalization -Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. -For example, one can use these techniques to generate images of themselves, using as little as one reference image. -These include teaching the model a new concept both in training free manner or through fine-tuning. - +Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. ## Inference From 9b792c1bb97b6fa2cb7a70972b3e11ef528983d9 Mon Sep 17 00:00:00 2001 From: Linoy Tsaban <57615435+linoytsaban@users.noreply.github.com> Date: Fri, 6 Sep 2024 13:54:54 +0300 Subject: [PATCH 10/23] Update packages/tasks/src/tasks/text-to-image/about.md Co-authored-by: Merve Noyan --- packages/tasks/src/tasks/text-to-image/about.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 3078eda0d0..54633a20b8 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -16,7 +16,7 @@ Different patterns can be generated to obtain unique pieces of fashion. Text-to- Architects can utilise the models to construct an environment based out on the requirements of the floor plan. This can also include the furniture that has to be placed in that environment. -##Task Variants +## Task Variants ### Image Editing From 39bdbe144e2d1f8772e2d6f525c32935ee577881 Mon Sep 17 00:00:00 2001 From: Linoy Tsaban <57615435+linoytsaban@users.noreply.github.com> Date: Fri, 6 Sep 2024 13:57:40 +0300 Subject: [PATCH 11/23] Update packages/tasks/src/tasks/text-to-image/about.md Co-authored-by: Merve Noyan --- packages/tasks/src/tasks/text-to-image/about.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 54633a20b8..ebe5038456 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -22,7 +22,7 @@ Architects can utilise the models to construct an environment based out on the r Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. -- Synthetic image editing: using text-to-image models to make adjustments to images that were initially created using an input prompt, while preserving the overall meaning or context of the original image. +- **Synthetic image editing:** Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) From 6a715782707b0f92e53f6eb054a6149bfab30f06 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Fri, 6 Sep 2024 14:51:45 +0300 Subject: [PATCH 12/23] simplify real image editing --- packages/tasks/src/tasks/text-to-image/about.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 3078eda0d0..295b36c2fd 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -26,13 +26,15 @@ Image editing with text-to-image models involves using text prompts to describe ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) -- Real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex, as it involves first obtaining a latent representation of the image, in the latent domain of the model that it can then manipulate. +- Real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) ### Personalization -Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. +Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. + +For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. ## Inference From 8c292c3bef6a7440fcce93cc98c5ed9cb703aa55 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Fri, 6 Sep 2024 14:53:31 +0300 Subject: [PATCH 13/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 12b75f4b39..c1e4b9a934 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -22,11 +22,11 @@ Architects can utilise the models to construct an environment based out on the r Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. -- **Synthetic image editing:** Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. +- **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) -- Real image editing: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. +- **Real image editing**: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) From e15a303091c38608c19be7ec4d047879c34e9c77 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Fri, 6 Sep 2024 14:54:11 +0300 Subject: [PATCH 14/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index c1e4b9a934..43f3369fdd 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -32,7 +32,7 @@ Image editing with text-to-image models involves using text prompts to describe ### Personalization -Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. +Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. From 1f9d0af6e24aee762607e8a1094f8497e860502a Mon Sep 17 00:00:00 2001 From: Linoy Tsaban <57615435+linoytsaban@users.noreply.github.com> Date: Sat, 7 Sep 2024 15:07:21 +0300 Subject: [PATCH 15/23] Update packages/tasks/src/tasks/text-to-image/data.ts Co-authored-by: Pedro Cuenca --- packages/tasks/src/tasks/text-to-image/data.ts | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/data.ts b/packages/tasks/src/tasks/text-to-image/data.ts index d16a5806c0..0c76b6a99e 100644 --- a/packages/tasks/src/tasks/text-to-image/data.ts +++ b/packages/tasks/src/tasks/text-to-image/data.ts @@ -92,7 +92,7 @@ const taskData: TaskDataCustom = { }, ], summary: - "Text-to-image is the task of generating images from input text. These pipelines can also be used modify and edit images based on text prompts.", + "Text-to-image is the task of generating images from input text. These pipelines can also be used to modify and edit images based on text prompts.", widgetModels: ["black-forest-labs/FLUX.1-dev"], youtubeId: "", }; From dbbfd4ff51191f81a2362d2ddb211d94dcc1d595 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sat, 7 Sep 2024 15:11:54 +0300 Subject: [PATCH 16/23] changes to personalization variant --- packages/tasks/src/tasks/text-to-image/about.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 43f3369fdd..257b2fc417 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -32,9 +32,9 @@ Image editing with text-to-image models involves using text prompts to describe ### Personalization -Personalization refers to techniques used to customize text-to-image models, where we introduce new subjects/concepts to the model so that we can then use the model to generate new images of those subjects with a text prompt. +Personalization refers to techniques used to customize text-to-image models. We introduce new subjects or concepts to the model, which the model can then generate when we refer to them with a text prompt. -For example, one can use these techniques to generate images of themselves, using as little as one reference image. These include teaching the model a new concept both in training free manner or through fine-tuning. +For example, you can use these techniques to generate images of your dog in imaginary settings, after you have taught the model using a few reference images of the subject (or just one in some cases). Teaching the model a new concept can be achieved through fine-tuning, or by using training-free techniques. ## Inference From 8bc84ccfea8005e736b4ad37fd008813ef0ce979 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sat, 7 Sep 2024 15:14:25 +0300 Subject: [PATCH 17/23] fix images --- packages/tasks/src/tasks/text-to-image/about.md | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 257b2fc417..08d0fe8c29 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -24,12 +24,11 @@ Image editing with text-to-image models involves using text prompts to describe - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. - ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/b20ecaa3f61372174c854e09fc856fdcce6f8494/--/default/train/0/image/image.png?Expires=1725455983&Signature=ykj3EnAENI6goXc7qI2Toq~P8P5IdS1DqNbSfH8vhgrdwaJoGH2cUbXWRgVAndhrHvRjrTTcU3YOyoExnot7zEhauyUEcqr-evRHDmGgfar52uEmfLbLCtNAcRK9Q85QOifupIH-X9x3rBUM03B0RIkHuto6wwRBAHireqr7QcD8hYRaNzACXrTbt-U7wHosZS8R1pdc3FDt7fDc3Qwh8XL0YoJqAoK8X8JnZEXIWTfGnCpygPBDbseDlYEzegGKzClAUgigQbomUk733VNtB3ol396uYkHCcjqjtgdhtEfAWQz-xM4eAhHpI~YEn7RQqRjB0RD0bPd1nHRU0wGUqA__&Key-Pair-Id=K3EI6M078Z3AC3) + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) - **Real image editing**: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. - ![Examples](https://datasets-server.huggingface.co/assets/diffusers/diffusers-images-docs/--/default/train/1/image/image.jpg?Expires=1725453082&Signature=MOCeELTChydgLRZT9ws8owCraSVrdcm6c7Vlnsi23rJ1Ocigl6gjRtXwmjVDCKuG2fB6Hw0Tmn8ZR0M7FPiA2fXpSuPEW4iJMoeQNiNCtkSSjjDisDXbBSRXW1TXJ-Z2c~VoJ4lmmeUdFpyFZ9W~BlI6r2xQLltfU400XKPe~UgE-vJ~xr9ni8zZmyYt1kVtV9Et~EBzWCQkKc2DO9gI9HnEg9z2hxDHp8Bak0HBRARM4ObhRYxieWqO4hOg1HVk4LSt2E8emIuDmhPUU4v8L097yFcI4D6JeoyNNn0q6nKQZqAZIzwP8iiLqqhSv~mJsO7YGnQck1-bzA~gAiVMpg__&Key-Pair-Id=K3EI6M078Z3AC3) - + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) ### Personalization Personalization refers to techniques used to customize text-to-image models. We introduce new subjects or concepts to the model, which the model can then generate when we refer to them with a text prompt. From 9754f203b01c693d09381ca7cf85b68fb7c58791 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sat, 7 Sep 2024 15:16:30 +0300 Subject: [PATCH 18/23] format --- packages/tasks/src/tasks/text-to-image/about.md | 1 + 1 file changed, 1 insertion(+) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 08d0fe8c29..7d67863691 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -29,6 +29,7 @@ Image editing with text-to-image models involves using text prompts to describe - **Real image editing**: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) + ### Personalization Personalization refers to techniques used to customize text-to-image models. We introduce new subjects or concepts to the model, which the model can then generate when we refer to them with a text prompt. From 4be7f1283b2674b11255c347008613448aed927c Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sun, 8 Sep 2024 13:16:37 +0300 Subject: [PATCH 19/23] change to image editing description --- packages/tasks/src/tasks/text-to-image/about.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 7d67863691..b5be4fe99e 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -20,7 +20,7 @@ Architects can utilise the models to construct an environment based out on the r ### Image Editing -Image editing with text-to-image models involves using text prompts to describe the wanted changes in an image and then follow them. +Image editing with text-to-image models involves modifying an image following edit instructions provided in a text prompt. - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. From a8580265bc308c6ed3b05e079915e316dfb4f5d0 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sun, 8 Sep 2024 13:21:33 +0300 Subject: [PATCH 20/23] add figure refs --- packages/tasks/src/tasks/text-to-image/about.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index b5be4fe99e..8a13dd0c2f 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -25,10 +25,12 @@ Image editing with text-to-image models involves modifying an image following ed - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) +_figure taken from ["InstructPix2Pix: Learning to Follow Image Editing Instructions"](https://www.timothybrooks.com/instruct-pix2pix)_ - **Real image editing**: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) +_figure taken from ["Prompt-to-Prompt Image Editing with Cross-Attention Control"](https://prompt-to-prompt.github.io)_ ### Personalization From 144c4cadf0b6bbc1dc88f4c9c9b76fb1ada165f8 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sun, 8 Sep 2024 13:29:39 +0300 Subject: [PATCH 21/23] format --- .../tasks/src/tasks/text-to-image/about.md | 20 +++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 8a13dd0c2f..77f1ab1a3f 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -22,16 +22,16 @@ Architects can utilise the models to construct an environment based out on the r Image editing with text-to-image models involves modifying an image following edit instructions provided in a text prompt. -- **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. - - ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) -_figure taken from ["InstructPix2Pix: Learning to Follow Image Editing Instructions"](https://www.timothybrooks.com/instruct-pix2pix)_ - -- **Real image editing**: similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. - - ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) -_figure taken from ["Prompt-to-Prompt Image Editing with Cross-Attention Control"](https://prompt-to-prompt.github.io)_ - + - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. + + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) + _Figure taken from ["InstructPix2Pix: Learning to Follow Image Editing Instructions"](https://www.timothybrooks.com/instruct-pix2pix)_ + + - **Real image editing**: Similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. + + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) + _Figure taken from ["Prompt-to-Prompt Image Editing with Cross-Attention Control"](https://prompt-to-prompt.github.io)_ + ### Personalization Personalization refers to techniques used to customize text-to-image models. We introduce new subjects or concepts to the model, which the model can then generate when we refer to them with a text prompt. From 43c297c0fde1c0987b421a95caf6e50c81d46bea Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sun, 8 Sep 2024 13:32:30 +0300 Subject: [PATCH 22/23] format --- .../tasks/src/tasks/text-to-image/about.md | 22 +++++++++---------- 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index 77f1ab1a3f..fa7608ee6f 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -18,20 +18,18 @@ Architects can utilise the models to construct an environment based out on the r ## Task Variants -### Image Editing - Image editing with text-to-image models involves modifying an image following edit instructions provided in a text prompt. - - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. - - ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) - _Figure taken from ["InstructPix2Pix: Learning to Follow Image Editing Instructions"](https://www.timothybrooks.com/instruct-pix2pix)_ - - - **Real image editing**: Similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. - - ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) - _Figure taken from ["Prompt-to-Prompt Image Editing with Cross-Attention Control"](https://prompt-to-prompt.github.io)_ - +- **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image. + + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/edit_p2p.png) + _Figure taken from ["InstructPix2Pix: Learning to Follow Image Editing Instructions"](https://www.timothybrooks.com/instruct-pix2pix)_ + +- **Real image editing**: Similar to synthetic image editing, except we're using real photos/images. This task is usually more complex. + + ![Examples](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/pix2pix.jpeg) + _Figure taken from ["Prompt-to-Prompt Image Editing with Cross-Attention Control"](https://prompt-to-prompt.github.io)_ + ### Personalization Personalization refers to techniques used to customize text-to-image models. We introduce new subjects or concepts to the model, which the model can then generate when we refer to them with a text prompt. From 9c306b68e2697ac2236fb32b829b868404e6a8e2 Mon Sep 17 00:00:00 2001 From: linoytsaban Date: Sun, 8 Sep 2024 17:20:25 +0300 Subject: [PATCH 23/23] add back missing title --- packages/tasks/src/tasks/text-to-image/about.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/packages/tasks/src/tasks/text-to-image/about.md b/packages/tasks/src/tasks/text-to-image/about.md index fa7608ee6f..e64a86615e 100644 --- a/packages/tasks/src/tasks/text-to-image/about.md +++ b/packages/tasks/src/tasks/text-to-image/about.md @@ -18,6 +18,8 @@ Architects can utilise the models to construct an environment based out on the r ## Task Variants +### Image Editing + Image editing with text-to-image models involves modifying an image following edit instructions provided in a text prompt. - **Synthetic image editing**: Adjusting images that were initially created using an input prompt while preserving the overall meaning or context of the original image.