jumafernandez commited on
Commit
8fa8b31
·
verified ·
1 Parent(s): 5d169cf

publish trace-repro-todbert-lite-ar-s7

Browse files
trace-repro-todbert-lite-ar-s7/best/config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "input_dim": 768,
3
+ "hidden_dim": 768,
4
+ "output_dim": 768,
5
+ "num_layers": 6,
6
+ "num_heads": 8,
7
+ "dropout": 0.1,
8
+ "max_turns": 64,
9
+ "attention_mode": "autoregressive",
10
+ "use_speaker_embeddings": true,
11
+ "num_speakers": 4,
12
+ "layer_norm": true,
13
+ "ff_dim": 3072,
14
+ "activation": "gelu",
15
+ "output_residual": false,
16
+ "arch": "v2",
17
+ "head_mode": "tied_continuous",
18
+ "head_transform": true
19
+ }
trace-repro-todbert-lite-ar-s7/best/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbeacbae1d38e1fb270e44b5ed0e8f0acc02afae2d72de3f8036729d69885831
3
+ size 179808064
trace-repro-todbert-lite-ar-s7/config.json ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "input_dim": 768,
3
+ "hidden_dim": 768,
4
+ "output_dim": 768,
5
+ "num_layers": 6,
6
+ "num_heads": 8,
7
+ "dropout": 0.1,
8
+ "max_turns": 64,
9
+ "attention_mode": "autoregressive",
10
+ "use_speaker_embeddings": true,
11
+ "num_speakers": 4,
12
+ "layer_norm": true,
13
+ "ff_dim": 3072,
14
+ "activation": "gelu",
15
+ "output_residual": false,
16
+ "arch": "v2",
17
+ "head_mode": "tied_continuous",
18
+ "head_transform": true
19
+ }
trace-repro-todbert-lite-ar-s7/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbeacbae1d38e1fb270e44b5ed0e8f0acc02afae2d72de3f8036729d69885831
3
+ size 179808064
trace-repro-todbert-lite-ar-s7/trainlog.jsonl ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 1, "train_loss": 4.91187, "val_loss": 5.24049, "epoch_sec": 85.6}
2
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 2, "train_loss": 4.12507, "val_loss": 5.05737, "epoch_sec": 69.3}
3
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 3, "train_loss": 3.93034, "val_loss": 4.96606, "epoch_sec": 69.4}
4
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 4, "train_loss": 3.77811, "val_loss": 4.92398, "epoch_sec": 69.3}
5
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 5, "train_loss": 3.62546, "val_loss": 4.89903, "epoch_sec": 69.3}
6
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 6, "train_loss": 3.46769, "val_loss": 4.87296, "epoch_sec": 69.4}
7
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 7, "train_loss": 3.31089, "val_loss": 4.85676, "epoch_sec": 70.1}
8
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 8, "train_loss": 3.16955, "val_loss": 4.83041, "epoch_sec": 71.3}
9
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 9, "train_loss": 3.04533, "val_loss": 4.8405, "epoch_sec": 71.5}
10
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 10, "train_loss": 2.93742, "val_loss": 4.81789, "epoch_sec": 71.2}
11
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 11, "train_loss": 2.84222, "val_loss": 4.79014, "epoch_sec": 74.7}
12
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 12, "train_loss": 2.76069, "val_loss": 4.76435, "epoch_sec": 73.9}
13
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 13, "train_loss": 2.69233, "val_loss": 4.75662, "epoch_sec": 74.8}
14
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 14, "train_loss": 2.63896, "val_loss": 4.75875, "epoch_sec": 74.0}
15
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 15, "train_loss": 2.6058, "val_loss": 4.7396, "epoch_sec": 74.6}