From 88efb1908254d19ab9efba9202fb600f06f13bcc Mon Sep 17 00:00:00 2001 From: Dmitriy Pastushenkov Date: Tue, 17 Dec 2024 12:51:37 +0100 Subject: [PATCH 1/2] blogs cross-referencing for image generation (python part) --- samples/python/image_generation/README.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/samples/python/image_generation/README.md b/samples/python/image_generation/README.md index 3e53f40fc4..a34613936a 100644 --- a/samples/python/image_generation/README.md +++ b/samples/python/image_generation/README.md @@ -24,6 +24,12 @@ Users can change the sample code and play with the following generation paramete > OpenVINO GenAI is written in C++ and uses `CppStdGenerator` random generator in Image Generation pipelines, while Diffusers library uses `torch.Generator` underhood. > To have the same results with HuggingFace, pass manually created `torch.Generator(device='cpu').manual_seed(seed)` to Diffusers generation pipelines and `openvino_genai.TorchGenerator(seed)` to OpenVINO GenAI pipelines as value for `generator` kwarg. +## OpenVINO blogs about Image Generation with GenAI + +Please follow the following blogs to setup your first hands-on experience with samples. + +* [How to generate images locally on AI PC with OpenVINO GenAI API](https://medium.com/openvino-toolkit/how-to-generate-images-locally-on-ai-pc-with-openvino-genai-api-220d08370958) + ## Download and convert the models and tokenizers The `--upgrade-strategy eager` option is needed to ensure `optimum-intel` is upgraded to the latest version. From 4baaf886eccb64bb055c3193adc1f1dc90e16437 Mon Sep 17 00:00:00 2001 From: Dmitriy Pastushenkov Date: Tue, 17 Dec 2024 12:54:49 +0100 Subject: [PATCH 2/2] blogs cross-referencing for image generation (c++ part) blogs cross-referencing for image generation (c++ part) --- samples/cpp/image_generation/README.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/samples/cpp/image_generation/README.md b/samples/cpp/image_generation/README.md index f8dc21cc39..5b026e10d8 100644 --- a/samples/cpp/image_generation/README.md +++ b/samples/cpp/image_generation/README.md @@ -24,6 +24,12 @@ Users can change the sample code and play with the following generation paramete > [!NOTE] > Image generated with HuggingFace / Optimum Intel is not the same generated by this C++ sample: C++ random generation with MT19937 results differ from `numpy.random.randn()` and `diffusers.utils.randn_tensor` (uses `torch.Generator` inside). So, it's expected that image generated by Diffusers and C++ versions provide different images, because latent images are initialize differently. +## OpenVINO blogs about Image Generation with GenAI + +Please follow the following blogs to setup your first hands-on experience with samples. + +* [How to generate images locally on AI PC with OpenVINO GenAI API](https://medium.com/openvino-toolkit/how-to-generate-images-locally-on-ai-pc-with-openvino-genai-api-220d08370958) + ## Download and convert the models and tokenizers The `--upgrade-strategy eager` option is needed to ensure `optimum-intel` is upgraded to the latest version.