{"id":47842379,"url":"https://github.com/jamesdconley/embeddageddon-organized","last_synced_at":"2026-04-03T21:01:16.261Z","repository":{"id":319943685,"uuid":"1080195117","full_name":"JamesDConley/embeddageddon-organized","owner":"JamesDConley","description":null,"archived":false,"fork":false,"pushed_at":"2025-10-31T04:22:54.000Z","size":80,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-10-31T06:14:01.014Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/JamesDConley.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-10-21T02:25:41.000Z","updated_at":"2025-10-31T04:22:58.000Z","dependencies_parsed_at":"2025-10-21T04:33:57.084Z","dependency_job_id":null,"html_url":"https://github.com/JamesDConley/embeddageddon-organized","commit_stats":null,"previous_names":["jamesdconley/embeddageddon-organized"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/JamesDConley/embeddageddon-organized","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JamesDConley%2Fembeddageddon-organized","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JamesDConley%2Fembeddageddon-organized/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JamesDConley%2Fembeddageddon-organized/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JamesDConley%2Fembeddageddon-organized/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/JamesDConley","download_url":"https://codeload.github.com/JamesDConley/embeddageddon-organized/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/JamesDConley%2Fembeddageddon-organized/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":31377113,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-03T17:53:18.093Z","status":"ssl_error","status_checked_at":"2026-04-03T17:53:17.617Z","response_time":107,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":"2026-04-03T21:01:10.380Z","updated_at":"2026-04-03T21:01:16.228Z","avatar_url":"https://github.com/JamesDConley.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Embedageddon\nGood artists copy, great artists steal!\n\n## Pull Embeddings\n`python src/extract_embeddings.py`\n\n## Setup Dataset\n`python src/preprocess_embeddings.py --embedding_dir data/embedding_extraction/embedding_dicts --output_dir data/autoencoder/datasets/full`\n\n## Train Autoencoder\n\n\nTrain with dataset via\n`python src/ae_train_memmap.py --preprocessed_dir data/autoencoder/datasets/full --batch_size 512 --num_workers 2 --epochs 100 --device cuda --output_dir data/autoencoder/trained_models/epochs_100_$(date +%Y%m%d_%H%M%S)`\n\nYou can also use the `ae_train_preloaded.py` if you have enough memory to fit all of the dictionaries into RAM. Note this will happen per worker, and multiple workers are needed to keep up with a single RTX Pro 6000. Even with it all preloaded into memory.\n\nThe memmap script I also find to get poor GPU utilization. There's definitely optimization that can be done here feeding the data to the training code.\n\n# Generate Embeddageddon Embeddings\n`python src/generate_embeddings.py --embedding_dicts_dir data/embedding_extraction/embedding_dicts --encoder_model_path data/autoencoder/trained_models/epochs_100_20251019_160346/models/embeddageddon_model_final.pth --output_dir data/embeddageddon_embeddings/xl --bottleneck_dim 7168`\n\n`python src/generate_embeddings.py --embedding_dicts_dir data/embedding_extraction/embedding_dicts --encoder_model_path data/autoencoder/trained_models/no_dropout_epochs_100_20251021_092845/models/embeddageddon_model_final.pth --output_dir data/embeddageddon_embeddings/no_dropout --bottleneck_dim 7168`\n\n`python src/generate_embeddings.py --embedding_dicts_dir data/embedding_extraction/embedding_dicts --encoder_model_path data/autoencoder/trained_models/no_dropout_epochs_100_20251021_092845/checkpoints/embeddageddon_model_epoch_50.pth --output_dir data/embeddageddon_embeddings/50_epochs_no_dropout --bottleneck_dim 7168`\n\n\n\n# Setup a Training Dataset\n`python src/generate_dataset.py --output_dir data/llm_datasets/redpajama_small`\n\n# Parameter Counts\n\nS - 298_447_744\nM - 1_147_412_224\nL - 5_597_896_192\nXL - 24_173_026_304\n\n# Chinchilla Token Counts\nS  - 5_968_954_880\nM  - 22_948_244_480\nL  - 111_957_923_840\nXL - 483_460_526_080","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjamesdconley%2Fembeddageddon-organized","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjamesdconley%2Fembeddageddon-organized","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjamesdconley%2Fembeddageddon-organized/lists"}