jumafernandez commited on
Commit
a4367a1
·
verified ·
1 Parent(s): 4be68dd

publish trace-repro-lite-ar-s7

Browse files
trace-repro-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-lite-ar-s7/best/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a0e47128b74166d68167554459e95a61f6090117fc8567a8d28293541632695
3
+ size 179808064
trace-repro-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-lite-ar-s7/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffc8f408c1ade31669ae69616fceb16297aabdb310554c6d4ddc7fdeb439cd27
3
+ size 179808064
trace-repro-lite-ar-s7/trainlog.jsonl ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 1, "train_loss": 4.24648, "val_loss": 5.05671, "epoch_sec": 139.7}
2
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 2, "train_loss": 3.314, "val_loss": 4.79532, "epoch_sec": 122.7}
3
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 3, "train_loss": 2.98461, "val_loss": 4.66298, "epoch_sec": 122.4}
4
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 4, "train_loss": 2.64425, "val_loss": 4.5695, "epoch_sec": 122.1}
5
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 5, "train_loss": 2.3228, "val_loss": 4.5444, "epoch_sec": 122.2}
6
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 6, "train_loss": 2.04286, "val_loss": 4.53061, "epoch_sec": 122.3}
7
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 7, "train_loss": 1.81093, "val_loss": 4.53516, "epoch_sec": 122.4}
8
+ {"model": "lite", "mode": "ar", "seed": 7, "lr": 0.0002, "canary": false, "init": "scratch", "epoch": 8, "train_loss": 1.62241, "val_loss": 4.55716, "epoch_sec": 122.6}