{
  "model_type": "memory_clip_seq",
  "auto_map": {
    "AutoConfig": "memory_clip_seq.MemoryCLIPSeqConfig",
    "AutoModel": "memory_clip_seq.MemoryCLIPSeqModel"
  },
  "architectures": [
    "MemoryCLIPSeqModel"
  ],
  "clip_model": "openai/clip-vit-large-patch14",
  "clip_hidden": 768,
  "clip_layers": 12,
  "clip_max_tokens": 77,
  "freeze_clip": true,
  "n_memory_tokens": 8,
  "bank_size": 64,
  "anchor_dim": 768,
  "n_bank_heads": 8,
  "bank_cross_layers": 2,
  "gate_type": "gru",
  "extract_layers": [
    1,
    3,
    5,
    7,
    9,
    11
  ],
  "layer_fusion": "learned",
  "max_content_tokens": 18,
  "segment_overlap": 4,
  "max_segments": 32,
  "cv_target": 0.2,
  "sequence_output": true,
  "sequence_len": 77,
  "sequence_recon_layers": 2,
  "sequence_recon_heads": 8,
  "collect_content_tokens": true,
  "max_content_positions": 256,
  "teacher_model": "answerdotai/ModernBERT-large",
  "teacher_hidden": 1024,
  "torch_dtype": "float32",
  "transformers_version": "4.48.0"
}