AbstractPhil commited on
Commit
e77da8c
·
verified ·
1 Parent(s): 27dbbd6

Upload folder using huggingface_hub

Browse files
beeper_rose_final.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee0f8495abba954d7125d2cb282d32b18465a3fb5d66550716b6e16ed52aa900
3
+ size 111301146
beeper_rose_final.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b8b682beba4d8e3a7ad2a4977bc429c7f83120fd2707879d85cd704030c68a2
3
+ size 111286288
config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "Beeper",
3
+ "variant": "rose",
4
+ "dataset_name": "roneneldan/TinyStories",
5
+ "dataset_split": "train[:10%]",
6
+ "context": 512,
7
+ "vocab_size": 8192,
8
+ "add_bos_eos": true,
9
+ "val_ratio": 0.01,
10
+ "test_ratio": 0.01,
11
+ "dim": 512,
12
+ "n_layers": 6,
13
+ "n_heads": 8,
14
+ "mlp_ratio": 4.0,
15
+ "dropout": 0.0,
16
+ "resid_dropout": 0.1,
17
+ "grad_checkpoint": true,
18
+ "compile_model": false,
19
+ "batch_size": 32,
20
+ "grad_accum_steps": 1,
21
+ "epochs": 3,
22
+ "lr": 0.0003,
23
+ "betas": [
24
+ 0.9,
25
+ 0.95
26
+ ],
27
+ "weight_decay": 0.1,
28
+ "warmup_steps": 500,
29
+ "max_steps": null,
30
+ "clip_grad": 1.0,
31
+ "min_lr": 1e-06,
32
+ "label_smoothing": 0.0,
33
+ "use_rose": true,
34
+ "lambda_rose": 0.1,
35
+ "rose_hidden_pool": "mean",
36
+ "use_pent": true,
37
+ "lambda_pent": 0.01,
38
+ "pent_align_weight": 0.5,
39
+ "pent_regular_weight": 0.5,
40
+ "mixed_precision": "bf16",
41
+ "log_dir": "./runs/beeper",
42
+ "log_interval": 50,
43
+ "ckpt_dir": "./beeper_checkpoints",
44
+ "export_dir": "./beeper_export",
45
+ "temperature": 0.9,
46
+ "top_k": 40,
47
+ "top_p": 0.9,
48
+ "repetition_penalty": 1.1,
49
+ "presence_penalty": 0.6,
50
+ "frequency_penalty": 0.0,
51
+ "upload_to_hub": true,
52
+ "hf_repo": "AbstractPhil/beeper-rose-tinystories-6l-512d-ctx512",
53
+ "tokenizer_path": "beeper.tokenizer.json",
54
+ "seed": 1337
55
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff