jumafernandez commited on
Commit
c6b71c1
·
verified ·
1 Parent(s): 82c76b3

publish trace-repro-deep-ar-s7

Browse files
trace-repro-deep-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": 12,
6
+ "num_heads": 12,
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-deep-ar-s7/best/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8765a6937ea3da93e76c13ff621608402509cf92da083019c8b73b539381f11
3
+ size 349928328
trace-repro-deep-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": 12,
6
+ "num_heads": 12,
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-deep-ar-s7/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:530763b77940ffc955ee13534287ee4c47cf3b731487b5d0c74b608decd1b777
3
+ size 349928328
trace-repro-deep-ar-s7/trainlog.jsonl ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 1, "train_loss": 4.38181, "val_loss": 5.12972, "epoch_sec": 224.4}
2
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 2, "train_loss": 3.37458, "val_loss": 4.85211, "epoch_sec": 199.6}
3
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 3, "train_loss": 3.01896, "val_loss": 4.70678, "epoch_sec": 199.1}
4
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 4, "train_loss": 2.67617, "val_loss": 4.63512, "epoch_sec": 198.9}
5
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 5, "train_loss": 2.34406, "val_loss": 4.61923, "epoch_sec": 199.2}
6
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 6, "train_loss": 2.049, "val_loss": 4.63363, "epoch_sec": 199.6}
7
+ {"model": "deep", "mode": "ar", "seed": 7, "lr": 0.0001, "canary": false, "init": "scratch", "epoch": 7, "train_loss": 1.80699, "val_loss": 4.67305, "epoch_sec": 199.6}