avilama_config.json

{
  "model_cfg": {
    "embed_dim": 256,
    "seed": 42,
    "max_length": 200,
    "flip_ratio": 0.5,
    "tensor_ratio": 1.0,
    "batch_size": 4,
    "num_workers": 4,
    "num_projection_layers": 1,
    "dropout": 0.1,
    "temperature": 1.0,
    "head_lr": 0.001,
    "weight_decay": 0.001,
    "patience": 1,
    "factor": 0.8,
    "epochs": 2,
    "vision_cfg": {
      "image_size": 336,
      "patch_size": 14,
      "image_embedding": 1024,
      "image_encoder": "vit_large_patch14_clip_336",
      "image_encoder_lr": 0.0001,
      "imgs_pixels": 255.0
    },
    "text_cfg": {
      "context_length": 512,
      "vocab_size": 501153,
      "text_embedding": 768,
      "text_encoder": "sartifyllc/AviLaBSE",
      "text_encoder_lr": 1e-05,
      "text_tokenizer": "sartifyllc/AviLaBSE"
    }
  },
  "preprocess_cfg": {
    "mean": [
      0.48145466,
      0.4578275,
      0.40821073
    ],
    "std": [
      0.26862954,
      0.26130258,
      0.27577711
    ]
  }
}