_target_: text_recognizer.network.mammut.MaMMUT
encoder:
  _target_: text_recognizer.network.cvit.CVit
  image_height: 7
  image_width: 128
  patch_height: 7
  patch_width: 1
  dim: &dim 512
  encoder:
    _target_: text_recognizer.network.transformer.encoder.Encoder
    dim: *dim
    heads: 8
    dim_head: 64
    ff_mult: 4
    depth: 2
    dropout_rate: 0.5
    use_rotary_emb: true
    one_kv_head: true
  stem:
    _target_: text_recognizer.network.convnext.convnext.ConvNext
    dim: 16
    dim_mults: [2, 8, 32]
    depths: [2, 2, 4]
    attn: null
  channels: 512
image_attn_pool:
  _target_: text_recognizer.network.transformer.attention.Attention
  dim: *dim
  heads: 8
  causal: false
  dim_head: 64
  ff_mult: 4
  dropout_rate: 0.0
  use_flash: true
  norm_context: true
  use_rotary_emb: false
  one_kv_head: true
decoder:
  _target_: text_recognizer.network.transformer.decoder.Decoder
  dim: *dim
  ff_mult: 4
  heads: 8
  dim_head: 64
  depth: 6
  dropout_rate: 0.5
  one_kv_head: true
dim: *dim
dim_latent: *dim
num_tokens: 57
pad_index: 3
num_image_queries: 64