{"id":23343903,"url":"https://github.com/bv7dev/vae-lab","last_synced_at":"2025-04-07T16:24:16.921Z","repository":{"id":268301388,"uuid":"903722864","full_name":"bv7dev/vae-lab","owner":"bv7dev","description":"Experimenting with simple VAEs for image generation","archived":false,"fork":false,"pushed_at":"2025-01-06T12:12:32.000Z","size":24575,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-02-13T18:46:12.571Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/bv7dev.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-12-15T11:53:48.000Z","updated_at":"2025-01-06T12:12:36.000Z","dependencies_parsed_at":"2024-12-15T22:27:18.460Z","dependency_job_id":"29db2802-7733-4374-a756-28d4efe44b23","html_url":"https://github.com/bv7dev/vae-lab","commit_stats":null,"previous_names":["bv7dev/vae-lab"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bv7dev%2Fvae-lab","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bv7dev%2Fvae-lab/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bv7dev%2Fvae-lab/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bv7dev%2Fvae-lab/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bv7dev","download_url":"https://codeload.github.com/bv7dev/vae-lab/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247685931,"owners_count":20979150,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-12-21T06:17:10.151Z","updated_at":"2025-04-07T16:24:16.907Z","avatar_url":"https://github.com/bv7dev.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# vae-lab\n\nExperiments with simple Variational-Autoencoder (VAE) in PyTorch\n\n1. create a Python venv\n2. install PyTorch\n3. explore scripts \u0026 notebook\n\n```python\n# Explore the latent space\n\nwith torch.no_grad():\n    dd = d[2] # desired direction\n\n    iter = torch.arange(10, device=device)\n    x = t + 50*iter[:, None]*dd[None, :]\n\n    path = vae.decode(x)\n    util.plot_image_grid(1, 10, path.cpu())\n```\n\n![continuous interpolation](public/continuous_interpolation.png)\n\n```python\n# Decode random latent vectors\n\nwith torch.no_grad():\n    z = torch.randn(10, LATENT_SIZE).to(device)\n    recon = vae.decode(z*3)\n    util.plot_image_grid(2, 5, recon.cpu())\n```\n\n![random latent sampling example](public/random_latent_sampling.png)\n\n#### ideas\n\n- try a round gaussian mask for original and reconstruction to give the center\n  of the image, the face, most weight for mse loss (fade into black, transparency?)\n\n![target of some experiment i dont know anymore :D](target.png)","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbv7dev%2Fvae-lab","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbv7dev%2Fvae-lab","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbv7dev%2Fvae-lab/lists"}