jumafernandez commited on
Commit
d205eab
·
verified ·
1 Parent(s): b6ec094

publish trace-repro-lite-bidi-s7

Browse files
trace-repro-lite-bidi-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": "bidirectional",
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-lite-bidi-s7/best/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a8115272140f167365457414a0da5270f48bb1a75d329a9af1d5eb507c79b55
3
+ size 179808064
trace-repro-lite-bidi-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": "bidirectional",
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-lite-bidi-s7/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2d26917105d0a0e61c59769bebcb355634880c30616f418cfacc60d0f64a493
3
+ size 179808064
trace-repro-lite-bidi-s7/trainlog.jsonl ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 1, "train_loss": 3.36376, "val_loss": 3.96003, "epoch_sec": 139.5}
2
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 2, "train_loss": 2.38804, "val_loss": 3.40447, "epoch_sec": 122.4}
3
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 3, "train_loss": 1.88534, "val_loss": 3.15094, "epoch_sec": 122.1}
4
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 4, "train_loss": 1.53577, "val_loss": 3.06706, "epoch_sec": 122.2}
5
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 5, "train_loss": 1.26011, "val_loss": 3.01988, "epoch_sec": 122.1}
6
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 6, "train_loss": 1.03596, "val_loss": 3.03178, "epoch_sec": 122.1}
7
+ {"model": "lite", "mode": "bidi", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 7, "train_loss": 0.85292, "val_loss": 3.04342, "epoch_sec": 122.3}