word2vec-2012/config.json

{
  "model_type": "word2vec",
  "architecture": "skip-gram",
  "embedding_dim": 300,
  "window_size": 15,
  "min_count": 30,
  "max_vocab_size": 50000,
  "negative_samples": 15,
  "epochs": 20,
  "training_data": "FineWeb dataset (filtered by year)",
  "language": "en",
  "license": "mit",
  "tags": [
    "word2vec",
    "embeddings",
    "yearly",
    "language-evolution",
    "fineweb"
  ]
}