publish trace-repro-deep-ar-s123-xlc

trace-repro-deep-ar-s123-xlc/best/config.json

@@ -0,0 +1,19 @@
+{
+  "input_dim": 768,
+  "hidden_dim": 768,
+  "output_dim": 768,
+  "num_layers": 12,
+  "num_heads": 12,
+  "dropout": 0.1,
+  "max_turns": 64,
+  "attention_mode": "autoregressive",
+  "use_speaker_embeddings": true,
+  "num_speakers": 4,
+  "layer_norm": true,
+  "ff_dim": 3072,
+  "activation": "gelu",
+  "output_residual": true,
+  "arch": "v2",
+  "head_mode": "tied_continuous",
+  "head_transform": true
+}

trace-repro-deep-ar-s123-xlc/best/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c7f34cfbff08a4c66647663378067ee2d5eb3afd84b5a36736eeedca22a833c
+size 349928328

trace-repro-deep-ar-s123-xlc/config.json

@@ -0,0 +1,19 @@
+{
+  "input_dim": 768,
+  "hidden_dim": 768,
+  "output_dim": 768,
+  "num_layers": 12,
+  "num_heads": 12,
+  "dropout": 0.1,
+  "max_turns": 64,
+  "attention_mode": "autoregressive",
+  "use_speaker_embeddings": true,
+  "num_speakers": 4,
+  "layer_norm": true,
+  "ff_dim": 3072,
+  "activation": "gelu",
+  "output_residual": true,
+  "arch": "v2",
+  "head_mode": "tied_continuous",
+  "head_transform": true
+}

trace-repro-deep-ar-s123-xlc/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5868054f4df2ed9a7534409c02663c61a45099909799ebc18e459837d4023ddd
+size 349928328

trace-repro-deep-ar-s123-xlc/trainlog.jsonl

@@ -0,0 +1,6 @@
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 1, "train_loss": 3.77393, "val_loss": 4.84943, "epoch_sec": 250.9}
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 2, "train_loss": 3.1671, "val_loss": 4.64582, "epoch_sec": 230.6}
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 3, "train_loss": 2.73129, "val_loss": 4.53625, "epoch_sec": 231.3}
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 4, "train_loss": 2.33631, "val_loss": 4.52618, "epoch_sec": 231.1}
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 5, "train_loss": 2.0082, "val_loss": 4.55654, "epoch_sec": 231.7}
+{"model": "deep", "mode": "ar", "seed": 123, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed123-repro/best", "epoch": 6, "train_loss": 1.75295, "val_loss": 4.61075, "epoch_sec": 229.4}