publish trace-repro-deep-ar-s7-xlc

trace-repro-deep-ar-s7-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-s7-xlc/best/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed74b4ca82735f9f46dfc5ea64f4920a3e83818ce1b9c31185bf3d31f6734c59
+size 349928328

trace-repro-deep-ar-s7-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-s7-xlc/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d0050ad3e376791a3341602edcbb47cae51b266c2ffdb66639409d5a82e9c65
+size 349928328

trace-repro-deep-ar-s7-xlc/trainlog.jsonl

@@ -0,0 +1,6 @@
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 1, "train_loss": 3.77419, "val_loss": 4.85936, "epoch_sec": 301.0}
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 2, "train_loss": 3.16661, "val_loss": 4.6436, "epoch_sec": 198.1}
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 3, "train_loss": 2.73102, "val_loss": 4.53133, "epoch_sec": 197.8}
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 4, "train_loss": 2.33744, "val_loss": 4.52593, "epoch_sec": 197.9}
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 5, "train_loss": 2.00927, "val_loss": 4.57414, "epoch_sec": 198.1}
+{"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "/home/jfernandez/trace/models/contextual-turn-encoder-base-deep-ar-xlc-tm-seed7-repro/best", "epoch": 6, "train_loss": 1.75367, "val_loss": 4.5808, "epoch_sec": 197.5}