jumafernandez commited on
Commit
e6ba9c5
·
verified ·
1 Parent(s): e54cbaf

publish trace-repro-deep-bidi-s7

Browse files
trace-repro-deep-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": 12,
6
+ "num_heads": 12,
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-deep-bidi-s7/best/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aba82038c592b48bd7af08cb9dfe7936b2a76814b1dcd88740225cf9c71d19af
3
+ size 349928328
trace-repro-deep-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": 12,
6
+ "num_heads": 12,
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-deep-bidi-s7/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d180de2202000a54e24d92fa1cfcdad8a15243a0561110112af62bad83313409
3
+ size 349928328
trace-repro-deep-bidi-s7/trainlog.jsonl ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 1, "train_loss": 3.49905, "val_loss": 4.06647, "epoch_sec": 224.4}
2
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 2, "train_loss": 2.49574, "val_loss": 3.56049, "epoch_sec": 196.2}
3
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 3, "train_loss": 1.97676, "val_loss": 3.29954, "epoch_sec": 196.0}
4
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 4, "train_loss": 1.5719, "val_loss": 3.20055, "epoch_sec": 195.8}
5
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 5, "train_loss": 1.2469, "val_loss": 3.19241, "epoch_sec": 195.8}
6
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 6, "train_loss": 0.99126, "val_loss": 3.17093, "epoch_sec": 196.0}
7
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 7, "train_loss": 0.79156, "val_loss": 3.20201, "epoch_sec": 195.9}
8
+ {"model": "deep", "mode": "bidi", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 8, "train_loss": 0.64166, "val_loss": 3.19889, "epoch_sec": 196.3}