add trace-dialog-lite-bidi-xlc-s1u

trace-dialog-lite-bidi-xlc-s1u/best/config.json

@@ -0,0 +1,19 @@
+{
+  "input_dim": 768,
+  "hidden_dim": 768,
+  "output_dim": 768,
+  "num_layers": 6,
+  "num_heads": 8,
+  "dropout": 0.1,
+  "max_turns": 64,
+  "attention_mode": "bidirectional",
+  "use_speaker_embeddings": true,
+  "num_speakers": 4,
+  "layer_norm": true,
+  "ff_dim": 3072,
+  "activation": "gelu",
+  "output_residual": false,
+  "arch": "v2",
+  "head_mode": "tied_continuous",
+  "head_transform": true
+}

trace-dialog-lite-bidi-xlc-s1u/best/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ebd140780d3244b1ebd49d36dd46d3b56ecdf6f5c64914f5410be99700a2d13
+size 179814312

trace-dialog-lite-bidi-xlc-s1u/best/training_args.json

@@ -0,0 +1,15 @@
+{
+  "tag": "xlc-s1u",
+  "recipe": "lite",
+  "stage": "s1u",
+  "corpus": "ultrachat_dialogs.slim.pkl",
+  "n_train_turns": 11072595,
+  "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best",
+  "n_layers": 6,
+  "n_heads": 8,
+  "lr": 0.0001,
+  "epoch": 6,
+  "train_loss": 1.43678,
+  "val_loss": 1.80021,
+  "epoch_sec": 1726.3
+}

trace-dialog-lite-bidi-xlc-s1u/config.json

@@ -0,0 +1,19 @@
+{
+  "input_dim": 768,
+  "hidden_dim": 768,
+  "output_dim": 768,
+  "num_layers": 6,
+  "num_heads": 8,
+  "dropout": 0.1,
+  "max_turns": 64,
+  "attention_mode": "bidirectional",
+  "use_speaker_embeddings": true,
+  "num_speakers": 4,
+  "layer_norm": true,
+  "ff_dim": 3072,
+  "activation": "gelu",
+  "output_residual": false,
+  "arch": "v2",
+  "head_mode": "tied_continuous",
+  "head_transform": true
+}

trace-dialog-lite-bidi-xlc-s1u/config.yaml

@@ -0,0 +1,66 @@
+model:
+  input_dim: 768
+  hidden_dim: 768
+  output_dim: 768
+  num_layers: 6
+  num_heads: 8
+  dropout: 0.1
+  max_turns: 64
+  attention_mode: bidirectional
+  use_speaker_embeddings: true
+  num_speakers: 4
+  layer_norm: true
+  ff_dim: 3072
+  activation: gelu
+  output_residual: false
+  arch: v2
+  head_mode: tied_continuous
+  head_transform: true
+losses:
+  masked_reconstruction:
+    enabled: null
+    mask_prob: 0.15
+    weight: 1.0
+  next_turn_prediction:
+    enabled: null
+    weight: 1.0
+  embedding_retrieval:
+    enabled: true
+    weight: 1.0
+    temperature: 0.07
+    normalize: true
+    candidate_mode: in_batch
+    target: auto
+  lambda_cosine: 1.0
+training:
+  seed: 42
+  batch_size: 128
+  epochs: 10
+  learning_rate: 0.0001
+  weight_decay: 0.01
+  warmup_ratio: 0.05
+  gradient_clip_norm: 1.0
+  device: mps
+  mixed_precision: false
+  num_workers: 0
+  log_interval: 100
+  output_dir: models/contextual-turn-encoder-base
+data:
+  path: null
+  max_turns: 64
+  window: truncate
+  stride: 32
+  dialogue_id_col: dialogue_id
+  turn_id_col: turn_id
+  utterance_col: utterance
+  speaker_col: speaker
+  embedding_col: embedding
+  speaker_map: null
+base_encoder:
+  model_name: sergioburdisso/dialog2flow-joint-bert-base
+  backend: sentence_transformers
+  batch_size: 64
+  normalize: false
+  freeze: true
+  device: auto
+  cache_dir: null

trace-dialog-lite-bidi-xlc-s1u/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0dee320a919d60f2c84b9a05d0c5d0e8214c6b59b352323f896df4141f44e1a
+size 179814312

trace-dialog-lite-bidi-xlc-s1u/training_args.json

@@ -0,0 +1,15 @@
+{
+  "tag": "xlc-s1u",
+  "recipe": "lite",
+  "stage": "s1u",
+  "corpus": "ultrachat_dialogs.slim.pkl",
+  "n_train_turns": 11072595,
+  "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best",
+  "n_layers": 6,
+  "n_heads": 8,
+  "lr": 0.0001,
+  "epoch": 8,
+  "train_loss": 1.31238,
+  "val_loss": 1.80067,
+  "epoch_sec": 1734.7
+}

trace-dialog-lite-bidi-xlc-s1u/trainlog.jsonl

@@ -0,0 +1,8 @@
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 1, "train_loss": 2.38652, "val_loss": 2.07587, "epoch_sec": 1718.7}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 2, "train_loss": 1.90032, "val_loss": 1.92406, "epoch_sec": 1747.6}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 3, "train_loss": 1.72795, "val_loss": 1.85364, "epoch_sec": 1764.0}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 4, "train_loss": 1.61319, "val_loss": 1.81903, "epoch_sec": 1730.0}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 5, "train_loss": 1.51789, "val_loss": 1.80411, "epoch_sec": 1740.0}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 6, "train_loss": 1.43678, "val_loss": 1.80021, "epoch_sec": 1726.3}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 7, "train_loss": 1.36852, "val_loss": 1.80038, "epoch_sec": 1730.6}
+{"tag": "xlc-s1u", "recipe": "lite", "stage": "s1u", "corpus": "ultrachat_dialogs.slim.pkl", "n_train_turns": 11072595, "init_from": "packages/trace-dialog/models/trace-dialog-lite-bidi-xlc-s1/best", "n_layers": 6, "n_heads": 8, "lr": 0.0001, "epoch": 8, "train_loss": 1.31238, "val_loss": 1.80067, "epoch_sec": 1734.7}