Прощай, Midjourney! БЕСПЛАТНЫЕ Альтернативы

The paragraph discusses the cost of subscribing to 'mourne' at $30 per month and 'chat gpt' at $20 per month, which might be expensive for some. The speaker introduces three free image generators that they personally use and value, urging viewers to watch the video to the end to learn about their capabilities. The paragraph emphasizes the importance of ensuring these tools are indeed free and transitions into discussing the main feature of the service, which is working with faces to create various images. The service adapts and processes facial features to fit a given scene, resulting in unique and surprising images. The speaker shares their experience with their own photograph, describing how the service transforms the original image into something unique and predictable.

This paragraph explains the process of creating a personalized image model using the service. Users can upload up to 40 images for training their model, with an emphasis on selecting photos with different emotions and angles but with good quality and lighting. Proper lighting is critical for the model to accurately recognize facial contours and features. After uploading, users should name their model, and it's important to prepare the content well as the result can be subpar without it. The model usually takes between 10 to 30 minutes to train. Once trained, it appears in the list of custom models, and users can use it in the regular image generator by selecting their model and starting to build. The paragraph also advises on how to prepare a query for the model, suggesting not to include too many details and to let the model understand who to create based on the photos and the request.

The speaker discusses how to get free access to image generation services, suggesting liking the video for those who were not aware that such services could be used for free. The paragraph provides a brief history of the service 'dali', which initially was not available to the general public but later introduced a simplified version called 'dali Mini'. 'Dali Mini' gained attention for its amusing and imperfect images created through browser requests. In the beginning of 2022, 'dali' presented a version that stood out among its competitors due to high-quality images and support for Russian language requests. Access to 'dali' was initially limited and required being on a waiting list, but later it became available to everyone through their website and a bin, allowing free use. By the end of September 2022, 'dali' was made freely available through Bing, Creator, and a subscription in the 'chat gpt'. The paragraph concludes by highlighting the intuitive and accessible process of creating images with 'dali', which supports the Russian language and is more accessible to non-English speaking users.

The paragraph delves into the advanced features of 'dali', emphasizing its ability to correct results based on the initial request, adding flexibility to the creative process and allowing for more precise matches to expectations. It also mentions the advanced algorithms that ensure a variety of results, even when the same request is repeated, providing users with unique image variations for any purpose. The paragraph explains how 'dali' starts with a text request where users describe what they want, and then an internal model like 'gpt-4' analyzes and adapts the request to improve understanding. 'Dali' then generates an image embodying the description with high detail and creativity. The paragraph also discusses the use of 'bage Creator' and 'Chat gpt' for optimizing the request for the best results, and the importance of clear and concise requests. It concludes by mentioning the 'ocus' tool, which combines the functionality of 'stable diffusion' with a simple interface, allowing users to find a balance between the speed and quality of image creation.

This paragraph introduces 'ocus' as a tool that offers additional creative possibilities by allowing users to use any image as a basis for a request, including images created with 'stable diffusion' or any other picture found online or uploaded from a local device. Users can request a series of variations based on their image, specifying the degree of difference from the original or a desire to change certain details. There is also an option to upscale images, increasing their size while maintaining quality. 'Oculus' also allows for the mixing of several images to create a unique synthesized result, with control over which image dominates to achieve the desired visual effect. The paragraph concludes by stating that while 'ocus' is not unique, it is one of many interfaces for 'stable diffusion', and encourages users to explore and find tools that meet their specific requirements, emphasizing the importance of finding a reliable tool that can help realize bold creative ideas without unexpected surprises or complications.