chore: import upstream snapshot with attribution

decoding/GAD/fairseq/config/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (c) Facebook, Inc. and its affiliates.
+#
+# This source code is licensed under the MIT license found in the
+# LICENSE file in the root directory of this source tree.

decoding/GAD/fairseq/config/config.yaml

@@ -0,0 +1,18 @@
+# @package _group_
+
+hydra:
+  run:
+    dir: .
+
+defaults:
+    - task: null
+    - model: null
+    - criterion: cross_entropy
+    - optimizer: null
+    - lr_scheduler: fixed
+    - bpe: null
+    - tokenizer: null
+    - scoring: null
+    - generation: null
+    - common_eval: null
+    - eval_lm: null

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_baevski_gbw.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "relu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 512
+decoder_output_dim: 512
+decoder_input_dim: 512
+decoder_ffn_embed_dim: 4096
+decoder_layers: 12
+decoder_attention_heads: 16
+decoder_normalize_before: true
+no_decoder_final_norm: true
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_baevski_wiki103.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "relu"
+dropout: 0.3
+attention_dropout: 0.1
+activation_dropout: 0.1
+relu_dropout: 0.1
+decoder_embed_dim: 1024
+decoder_output_dim: 1024
+decoder_input_dim: 1024
+decoder_ffn_embed_dim: 4096
+decoder_layers: 16
+decoder_attention_heads: 8
+decoder_normalize_before: true
+no_decoder_final_norm: true
+adaptive_softmax_cutoff: "20000,60000"
+adaptive_softmax_dropout: 0.2
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: true
+adaptive_input_factor: 4
+adaptive_input_cutoff: "20000,60000"
+tie_adaptive_weights: true
+tie_adaptive_proj: true
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_big.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "relu"
+dropout: 0.1
+attention_dropout: 0.0
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 1024
+decoder_output_dim: 1024
+decoder_input_dim: 1024
+decoder_ffn_embed_dim: 4096
+decoder_layers: 12
+decoder_attention_heads: 16
+decoder_normalize_before: true
+no_decoder_final_norm: false
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_gbw.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "relu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 512
+decoder_output_dim: 512
+decoder_input_dim: 512
+decoder_ffn_embed_dim: 4096
+decoder_layers: 12
+decoder_attention_heads: 16
+decoder_normalize_before: true
+no_decoder_final_norm: true
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_gpt.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "gelu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 768
+decoder_output_dim: 768
+decoder_input_dim: 768
+decoder_ffn_embed_dim: 3072
+decoder_layers: 12
+decoder_attention_heads: 12
+decoder_normalize_before: true
+no_decoder_final_norm: false
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_gpt2_big.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "gelu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 1600
+decoder_output_dim: 1600
+decoder_input_dim: 1600
+decoder_ffn_embed_dim: 6400
+decoder_layers: 48
+decoder_attention_heads: 25
+decoder_normalize_before: true
+no_decoder_final_norm: false
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_gpt2_medium.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "gelu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 1280
+decoder_output_dim: 1280
+decoder_input_dim: 1280
+decoder_ffn_embed_dim: 5120
+decoder_layers: 36
+decoder_attention_heads: 20
+decoder_normalize_before: true
+no_decoder_final_norm: false
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_gpt2_small.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "gelu"
+dropout: 0.1
+attention_dropout: 0.1
+activation_dropout: 0.0
+relu_dropout: 0.0
+decoder_embed_dim: 1024
+decoder_output_dim: 1024
+decoder_input_dim: 1024
+decoder_ffn_embed_dim: 4096
+decoder_layers: 24
+decoder_attention_heads: 16
+decoder_normalize_before: true
+no_decoder_final_norm: false
+adaptive_softmax_cutoff: null
+adaptive_softmax_dropout: 0
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: false
+adaptive_input_factor: 4
+adaptive_input_cutoff: null
+tie_adaptive_weights: false
+tie_adaptive_proj: false
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/transformer_lm/transformer_lm_wiki103.yaml

@@ -0,0 +1,36 @@
+# @package _group_
+activation_fn: "relu"
+dropout: 0.3
+attention_dropout: 0.1
+activation_dropout: 0.1
+relu_dropout: 0.1
+decoder_embed_dim: 1024
+decoder_output_dim: 1024
+decoder_input_dim: 1024
+decoder_ffn_embed_dim: 4096
+decoder_layers: 16
+decoder_attention_heads: 8
+decoder_normalize_before: true
+no_decoder_final_norm: true
+adaptive_softmax_cutoff: "20000,60000"
+adaptive_softmax_dropout: 0.2
+adaptive_softmax_factor: 4
+no_token_positional_embeddings: false
+share_decoder_input_output_embed: false
+character_embeddings: false
+character_filters: "[(1, 64), (2, 128), (3, 192), (4, 256), (5, 256), (6, 256), (7, 256)]"
+character_embedding_dim: 4
+char_embedder_highway_layers: 2
+adaptive_input: true
+adaptive_input_factor: 4
+adaptive_input_cutoff: "20000,60000"
+tie_adaptive_weights: true
+tie_adaptive_proj: true
+decoder_learned_pos: false
+decoder_layerdrop: 0
+decoder_layers_to_keep: null
+layernorm_embedding: false
+no_scale_embedding: false
+quant_noise_pq: 0
+quant_noise_pq_block_size: 8
+quant_noise_scalar: 0

decoding/GAD/fairseq/config/model/wav2vec/vq_wav2vec_gumbel.yaml

@@ -0,0 +1,5 @@
+# @package _group_
+activation: gelu
+vq_type: gumbel
+vq_depth: 2
+combine_groups: true

decoding/GAD/fairseq/config/model/wav2vec2/wav2vec2_base.yaml

@@ -0,0 +1,8 @@
+# @package _group_
+
+quantize_targets: true
+final_dim: 256
+encoder_layerdrop: 0.05
+dropout_input: 0.1
+dropout_features: 0.1
+feature_grad_mult: 0.1

decoding/GAD/fairseq/config/model/wav2vec2/wav2vec2_large.yaml

@@ -0,0 +1,20 @@
+# @package _group_
+
+quantize_targets: true
+extractor_mode: layer_norm
+layer_norm_first: true
+final_dim: 768
+latent_temp: [2.0,0.1,0.999995]
+encoder_layerdrop: 0.0
+dropout_input: 0.0
+dropout_features: 0.0
+dropout: 0.0
+attention_dropout: 0.0
+conv_bias: true
+
+encoder_layers: 24
+encoder_embed_dim: 1024
+encoder_ffn_embed_dim: 4096
+encoder_attention_heads: 16
+
+feature_grad_mult: 1.0