Upload 4 files
Browse files- tokenizer.json +0 -0
- tokenizer_config.json +32 -0
- trainer_state.json +79 -0
- training_args.bin +3 -0
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"__type": "AddedToken",
|
4 |
+
"content": "<s>",
|
5 |
+
"lstrip": false,
|
6 |
+
"normalized": true,
|
7 |
+
"rstrip": false,
|
8 |
+
"single_word": false
|
9 |
+
},
|
10 |
+
"clean_up_tokenization_spaces": false,
|
11 |
+
"eos_token": {
|
12 |
+
"__type": "AddedToken",
|
13 |
+
"content": "</s>",
|
14 |
+
"lstrip": false,
|
15 |
+
"normalized": true,
|
16 |
+
"rstrip": false,
|
17 |
+
"single_word": false
|
18 |
+
},
|
19 |
+
"model_max_length": 1000000000000000019884624838656,
|
20 |
+
"pad_token": null,
|
21 |
+
"sp_model_kwargs": {},
|
22 |
+
"tokenizer_class": "LlamaTokenizer",
|
23 |
+
"unk_token": {
|
24 |
+
"__type": "AddedToken",
|
25 |
+
"content": "<unk>",
|
26 |
+
"lstrip": false,
|
27 |
+
"normalized": true,
|
28 |
+
"rstrip": false,
|
29 |
+
"single_word": false
|
30 |
+
},
|
31 |
+
"use_default_system_prompt": true
|
32 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,79 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 100.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 100,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 10.0,
|
13 |
+
"learning_rate": 0.0002,
|
14 |
+
"loss": 0.4335,
|
15 |
+
"step": 10
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 20.0,
|
19 |
+
"learning_rate": 0.0002,
|
20 |
+
"loss": 0.1164,
|
21 |
+
"step": 20
|
22 |
+
},
|
23 |
+
{
|
24 |
+
"epoch": 30.0,
|
25 |
+
"learning_rate": 0.0002,
|
26 |
+
"loss": 0.0365,
|
27 |
+
"step": 30
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"epoch": 40.0,
|
31 |
+
"learning_rate": 0.0002,
|
32 |
+
"loss": 0.0206,
|
33 |
+
"step": 40
|
34 |
+
},
|
35 |
+
{
|
36 |
+
"epoch": 50.0,
|
37 |
+
"learning_rate": 0.0002,
|
38 |
+
"loss": 0.0027,
|
39 |
+
"step": 50
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 60.0,
|
43 |
+
"learning_rate": 0.0002,
|
44 |
+
"loss": 0.0004,
|
45 |
+
"step": 60
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"epoch": 70.0,
|
49 |
+
"learning_rate": 0.0002,
|
50 |
+
"loss": 0.0001,
|
51 |
+
"step": 70
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 80.0,
|
55 |
+
"learning_rate": 0.0002,
|
56 |
+
"loss": 0.0,
|
57 |
+
"step": 80
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 90.0,
|
61 |
+
"learning_rate": 0.0002,
|
62 |
+
"loss": 0.0,
|
63 |
+
"step": 90
|
64 |
+
},
|
65 |
+
{
|
66 |
+
"epoch": 100.0,
|
67 |
+
"learning_rate": 0.0002,
|
68 |
+
"loss": 0.0,
|
69 |
+
"step": 100
|
70 |
+
}
|
71 |
+
],
|
72 |
+
"logging_steps": 10,
|
73 |
+
"max_steps": 100,
|
74 |
+
"num_train_epochs": 100,
|
75 |
+
"save_steps": 100,
|
76 |
+
"total_flos": 161297240064000.0,
|
77 |
+
"trial_name": null,
|
78 |
+
"trial_params": null
|
79 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:41a12b339d9535eac5561489bffc4d94f100a40a6c7956688658176fee56275a
|
3 |
+
size 4027
|