2 files changed, 0 insertions, 102 deletions
diff --git a/training/conf/network/conv_perceiver.yaml b/training/conf/network/conv_perceiver.yaml
deleted file mode 100644
index 2e12db9..0000000
--- a/training/conf/network/conv_perceiver.yaml
+++ /dev/null
@@ -1,37 +0,0 @@
-_target_: text_recognizer.networks.ConvPerceiver
-input_dims: [1, 1, 576, 640]
-hidden_dim: &hidden_dim 128
-num_classes: &num_classes 58
-max_length: &max_length 89
-num_queries: *max_length
-queries_dim: &queries_dim 64
-pad_index: 3
-encoder:
-  _target_: text_recognizer.networks.EfficientNet
-  arch: b0
-  stochastic_dropout_rate: 0.2
-  bn_momentum: 0.99
-  bn_eps: 1.0e-3
-  depth: 5
-  out_channels: *hidden_dim
-decoder:
-  _target_: text_recognizer.networks.perceiver.PerceiverIO
-  dim: 192
-  cross_heads: 1
-  cross_head_dim: 64
-  num_latents: 256
-  latent_dim: 512
-  latent_heads: 8
-  depth: 6
-  queries_dim: 128
-  logits_dim: *num_classes
-pixel_embedding:
-  _target_: text_recognizer.networks.transformer.embeddings.axial.AxialPositionalEmbeddingImage
-  dim: 64
-  axial_shape: [3, 64]
-  axial_dims: [32, 32]
-query_pos_emb:
-  _target_: text_recognizer.networks.transformer.embeddings.absolute.AbsolutePositionalEmbedding
-  dim: 64
-  max_seq_len: *max_length
-  l2norm_embed: true
diff --git a/training/conf/network/vq_transformer.yaml b/training/conf/network/vq_transformer.yaml
deleted file mode 100644
index d62a4b7..0000000
--- a/training/conf/network/vq_transformer.yaml
+++ /dev/null
@@ -1,65 +0,0 @@
-_target_: text_recognizer.networks.VqTransformer
-input_dims: [1, 1, 576, 640]
-hidden_dim: &hidden_dim 144
-num_classes: 58
-pad_index: 3
-encoder:
-  _target_: text_recognizer.networks.EfficientNet
-  arch: b0
-  stochastic_dropout_rate: 0.2
-  bn_momentum: 0.99
-  bn_eps: 1.0e-3
-  depth: 5
-  out_channels: *hidden_dim
-decoder:
-  _target_: text_recognizer.networks.transformer.Decoder
-  depth: 6
-  block:
-    _target_: text_recognizer.networks.transformer.DecoderBlock
-    self_attn:
-      _target_: text_recognizer.networks.transformer.Attention
-      dim: *hidden_dim
-      num_heads: 8
-      dim_head: 64
-      dropout_rate: &dropout_rate 0.4
-      causal: true
-      rotary_embedding:
-        _target_: text_recognizer.networks.transformer.RotaryEmbedding
-        dim: 64
-    cross_attn:
-      _target_: text_recognizer.networks.transformer.Attention
-      dim: *hidden_dim
-      num_heads: 8
-      dim_head: 64
-      dropout_rate: *dropout_rate
-      causal: false
-    norm:
-      _target_: text_recognizer.networks.transformer.RMSNorm
-      dim: *hidden_dim
-    ff:
-      _target_: text_recognizer.networks.transformer.FeedForward
-      dim: *hidden_dim
-      dim_out: null
-      expansion_factor: 2
-      glu: true
-      dropout_rate: *dropout_rate
-pixel_embedding:
-  _target_: text_recognizer.networks.transformer.AxialPositionalEmbedding
-  dim: *hidden_dim
-  shape: [18, 79]
-quantizer:
-  _target_: text_recognizer.networks.quantizer.VectorQuantizer
-  input_dim: *hidden_dim
-  codebook:
-    _target_: text_recognizer.networks.quantizer.CosineSimilarityCodebook
-    dim: 16
-    codebook_size: 64
-    kmeans_init: true
-    kmeans_iters: 10
-    decay: 0.8
-    eps: 1.0e-5
-    threshold_dead: 2
-    temperature: 0.0
-  commitment: 0.25
-  ort_reg_weight: 10
-  ort_reg_max_codes: 64