End of training
Browse files
README.md
CHANGED
@@ -1,11 +1,23 @@
|
|
1 |
---
|
2 |
tags:
|
3 |
- generated_from_trainer
|
|
|
|
|
4 |
metrics:
|
5 |
- accuracy
|
6 |
model-index:
|
7 |
- name: opt-babylm2-subset-default-3e-4
|
8 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
---
|
10 |
|
11 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -13,7 +25,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
13 |
|
14 |
# opt-babylm2-subset-default-3e-4
|
15 |
|
16 |
-
This model was trained from scratch on
|
17 |
It achieves the following results on the evaluation set:
|
18 |
- Loss: 2.3776
|
19 |
- Accuracy: 0.5327
|
|
|
1 |
---
|
2 |
tags:
|
3 |
- generated_from_trainer
|
4 |
+
datasets:
|
5 |
+
- kanishka/babylm2-subset
|
6 |
metrics:
|
7 |
- accuracy
|
8 |
model-index:
|
9 |
- name: opt-babylm2-subset-default-3e-4
|
10 |
+
results:
|
11 |
+
- task:
|
12 |
+
name: Causal Language Modeling
|
13 |
+
type: text-generation
|
14 |
+
dataset:
|
15 |
+
name: kanishka/babylm2-subset
|
16 |
+
type: kanishka/babylm2-subset
|
17 |
+
metrics:
|
18 |
+
- name: Accuracy
|
19 |
+
type: accuracy
|
20 |
+
value: 0.5327396962492645
|
21 |
---
|
22 |
|
23 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
25 |
|
26 |
# opt-babylm2-subset-default-3e-4
|
27 |
|
28 |
+
This model was trained from scratch on the kanishka/babylm2-subset dataset.
|
29 |
It achieves the following results on the evaluation set:
|
30 |
- Loss: 2.3776
|
31 |
- Accuracy: 0.5327
|
all_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"eval_accuracy": 0.5327396962492645,
|
4 |
+
"eval_loss": 2.377573251724243,
|
5 |
+
"eval_runtime": 124.4661,
|
6 |
+
"eval_samples": 46951,
|
7 |
+
"eval_samples_per_second": 377.219,
|
8 |
+
"eval_steps_per_second": 5.897,
|
9 |
+
"perplexity": 10.77871387443153,
|
10 |
+
"total_flos": 5.9695438005504e+17,
|
11 |
+
"train_loss": 2.237272139724321,
|
12 |
+
"train_runtime": 31036.7572,
|
13 |
+
"train_samples": 453383,
|
14 |
+
"train_samples_per_second": 146.079,
|
15 |
+
"train_steps_per_second": 4.565
|
16 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"eval_accuracy": 0.5327396962492645,
|
4 |
+
"eval_loss": 2.377573251724243,
|
5 |
+
"eval_runtime": 124.4661,
|
6 |
+
"eval_samples": 46951,
|
7 |
+
"eval_samples_per_second": 377.219,
|
8 |
+
"eval_steps_per_second": 5.897,
|
9 |
+
"perplexity": 10.77871387443153
|
10 |
+
}
|
runs/Jul24_22-53-38_phyl-ling-p01.la.utexas.edu/events.out.tfevents.1721911056.phyl-ling-p01.la.utexas.edu.132712.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f902d4566853072bec641ec0a18171a081d044cfdc6327673cbc1ce69376a2a9
|
3 |
+
size 417
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"total_flos": 5.9695438005504e+17,
|
4 |
+
"train_loss": 2.237272139724321,
|
5 |
+
"train_runtime": 31036.7572,
|
6 |
+
"train_samples": 453383,
|
7 |
+
"train_samples_per_second": 146.079,
|
8 |
+
"train_steps_per_second": 4.565
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,1119 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 10.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 141690,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.07057661091114405,
|
13 |
+
"grad_norm": 0.8544681072235107,
|
14 |
+
"learning_rate": 9.375e-06,
|
15 |
+
"loss": 5.8075,
|
16 |
+
"step": 1000
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.1411532218222881,
|
20 |
+
"grad_norm": 0.9348427057266235,
|
21 |
+
"learning_rate": 1.875e-05,
|
22 |
+
"loss": 3.9507,
|
23 |
+
"step": 2000
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.21172983273343213,
|
27 |
+
"grad_norm": 0.9987612366676331,
|
28 |
+
"learning_rate": 2.8125e-05,
|
29 |
+
"loss": 3.6154,
|
30 |
+
"step": 3000
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.2823064436445762,
|
34 |
+
"grad_norm": 0.9734600782394409,
|
35 |
+
"learning_rate": 3.75e-05,
|
36 |
+
"loss": 3.444,
|
37 |
+
"step": 4000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.3528830545557202,
|
41 |
+
"grad_norm": 0.9963154196739197,
|
42 |
+
"learning_rate": 4.6874999999999994e-05,
|
43 |
+
"loss": 3.2983,
|
44 |
+
"step": 5000
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.42345966546686425,
|
48 |
+
"grad_norm": 0.8833392858505249,
|
49 |
+
"learning_rate": 5.625e-05,
|
50 |
+
"loss": 3.1806,
|
51 |
+
"step": 6000
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.49403627637800834,
|
55 |
+
"grad_norm": 0.8395134806632996,
|
56 |
+
"learning_rate": 6.5625e-05,
|
57 |
+
"loss": 3.0762,
|
58 |
+
"step": 7000
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.5646128872891524,
|
62 |
+
"grad_norm": 0.9056613445281982,
|
63 |
+
"learning_rate": 7.5e-05,
|
64 |
+
"loss": 3.0103,
|
65 |
+
"step": 8000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.6351894982002965,
|
69 |
+
"grad_norm": 0.8697331547737122,
|
70 |
+
"learning_rate": 8.437499999999999e-05,
|
71 |
+
"loss": 2.9146,
|
72 |
+
"step": 9000
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.7057661091114404,
|
76 |
+
"grad_norm": 0.7743774056434631,
|
77 |
+
"learning_rate": 9.374999999999999e-05,
|
78 |
+
"loss": 2.8472,
|
79 |
+
"step": 10000
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.7763427200225845,
|
83 |
+
"grad_norm": 0.769959568977356,
|
84 |
+
"learning_rate": 0.00010312499999999999,
|
85 |
+
"loss": 2.8093,
|
86 |
+
"step": 11000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.8469193309337285,
|
90 |
+
"grad_norm": 0.7926977872848511,
|
91 |
+
"learning_rate": 0.0001125,
|
92 |
+
"loss": 2.7498,
|
93 |
+
"step": 12000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.9174959418448726,
|
97 |
+
"grad_norm": 0.7256486415863037,
|
98 |
+
"learning_rate": 0.000121865625,
|
99 |
+
"loss": 2.714,
|
100 |
+
"step": 13000
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.9880725527560167,
|
104 |
+
"grad_norm": 0.7287374138832092,
|
105 |
+
"learning_rate": 0.000131240625,
|
106 |
+
"loss": 2.6575,
|
107 |
+
"step": 14000
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 1.0,
|
111 |
+
"eval_accuracy": 0.4749594473317534,
|
112 |
+
"eval_loss": 2.8570854663848877,
|
113 |
+
"eval_runtime": 123.5266,
|
114 |
+
"eval_samples_per_second": 380.088,
|
115 |
+
"eval_steps_per_second": 5.942,
|
116 |
+
"step": 14169
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 1.0586491636671607,
|
120 |
+
"grad_norm": 0.7397704124450684,
|
121 |
+
"learning_rate": 0.00014060625,
|
122 |
+
"loss": 2.6238,
|
123 |
+
"step": 15000
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.1292257745783048,
|
127 |
+
"grad_norm": 0.7099004983901978,
|
128 |
+
"learning_rate": 0.000149971875,
|
129 |
+
"loss": 2.5914,
|
130 |
+
"step": 16000
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 1.1998023854894488,
|
134 |
+
"grad_norm": 0.6785891056060791,
|
135 |
+
"learning_rate": 0.000159346875,
|
136 |
+
"loss": 2.5765,
|
137 |
+
"step": 17000
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 1.2703789964005927,
|
141 |
+
"grad_norm": 0.6459276080131531,
|
142 |
+
"learning_rate": 0.000168703125,
|
143 |
+
"loss": 2.5484,
|
144 |
+
"step": 18000
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 1.340955607311737,
|
148 |
+
"grad_norm": 0.6296180486679077,
|
149 |
+
"learning_rate": 0.000178078125,
|
150 |
+
"loss": 2.5329,
|
151 |
+
"step": 19000
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 1.4115322182228809,
|
155 |
+
"grad_norm": 0.648757815361023,
|
156 |
+
"learning_rate": 0.00018745312499999998,
|
157 |
+
"loss": 2.5078,
|
158 |
+
"step": 20000
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 1.482108829134025,
|
162 |
+
"grad_norm": 0.6126940250396729,
|
163 |
+
"learning_rate": 0.00019681874999999998,
|
164 |
+
"loss": 2.5066,
|
165 |
+
"step": 21000
|
166 |
+
},
|
167 |
+
{
|
168 |
+
"epoch": 1.552685440045169,
|
169 |
+
"grad_norm": 0.5499350428581238,
|
170 |
+
"learning_rate": 0.00020618437499999995,
|
171 |
+
"loss": 2.4882,
|
172 |
+
"step": 22000
|
173 |
+
},
|
174 |
+
{
|
175 |
+
"epoch": 1.623262050956313,
|
176 |
+
"grad_norm": 0.7012745141983032,
|
177 |
+
"learning_rate": 0.00021555937499999998,
|
178 |
+
"loss": 2.4746,
|
179 |
+
"step": 23000
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 1.6938386618674572,
|
183 |
+
"grad_norm": 0.563831627368927,
|
184 |
+
"learning_rate": 0.00022493437499999998,
|
185 |
+
"loss": 2.4607,
|
186 |
+
"step": 24000
|
187 |
+
},
|
188 |
+
{
|
189 |
+
"epoch": 1.764415272778601,
|
190 |
+
"grad_norm": 0.4928041696548462,
|
191 |
+
"learning_rate": 0.00023430937499999997,
|
192 |
+
"loss": 2.4454,
|
193 |
+
"step": 25000
|
194 |
+
},
|
195 |
+
{
|
196 |
+
"epoch": 1.8349918836897452,
|
197 |
+
"grad_norm": 0.5389479398727417,
|
198 |
+
"learning_rate": 0.00024367499999999997,
|
199 |
+
"loss": 2.4429,
|
200 |
+
"step": 26000
|
201 |
+
},
|
202 |
+
{
|
203 |
+
"epoch": 1.9055684946008893,
|
204 |
+
"grad_norm": 0.549089252948761,
|
205 |
+
"learning_rate": 0.000253040625,
|
206 |
+
"loss": 2.4239,
|
207 |
+
"step": 27000
|
208 |
+
},
|
209 |
+
{
|
210 |
+
"epoch": 1.9761451055120332,
|
211 |
+
"grad_norm": 0.5027530193328857,
|
212 |
+
"learning_rate": 0.000262415625,
|
213 |
+
"loss": 2.4179,
|
214 |
+
"step": 28000
|
215 |
+
},
|
216 |
+
{
|
217 |
+
"epoch": 2.0,
|
218 |
+
"eval_accuracy": 0.4989820084802377,
|
219 |
+
"eval_loss": 2.6257801055908203,
|
220 |
+
"eval_runtime": 124.6794,
|
221 |
+
"eval_samples_per_second": 376.574,
|
222 |
+
"eval_steps_per_second": 5.887,
|
223 |
+
"step": 28338
|
224 |
+
},
|
225 |
+
{
|
226 |
+
"epoch": 2.0467217164231775,
|
227 |
+
"grad_norm": 0.4511743485927582,
|
228 |
+
"learning_rate": 0.000271790625,
|
229 |
+
"loss": 2.3885,
|
230 |
+
"step": 29000
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 2.1172983273343213,
|
234 |
+
"grad_norm": 0.4871021807193756,
|
235 |
+
"learning_rate": 0.00028115624999999994,
|
236 |
+
"loss": 2.378,
|
237 |
+
"step": 30000
|
238 |
+
},
|
239 |
+
{
|
240 |
+
"epoch": 2.1878749382454656,
|
241 |
+
"grad_norm": 0.4743562638759613,
|
242 |
+
"learning_rate": 0.00029053124999999994,
|
243 |
+
"loss": 2.3721,
|
244 |
+
"step": 31000
|
245 |
+
},
|
246 |
+
{
|
247 |
+
"epoch": 2.2584515491566095,
|
248 |
+
"grad_norm": 0.4668987989425659,
|
249 |
+
"learning_rate": 0.00029990624999999993,
|
250 |
+
"loss": 2.3584,
|
251 |
+
"step": 32000
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 2.3290281600677534,
|
255 |
+
"grad_norm": 0.4822200834751129,
|
256 |
+
"learning_rate": 0.0002972951043850852,
|
257 |
+
"loss": 2.3592,
|
258 |
+
"step": 33000
|
259 |
+
},
|
260 |
+
{
|
261 |
+
"epoch": 2.3996047709788977,
|
262 |
+
"grad_norm": 0.43820998072624207,
|
263 |
+
"learning_rate": 0.0002945601239857781,
|
264 |
+
"loss": 2.3459,
|
265 |
+
"step": 34000
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 2.4701813818900415,
|
269 |
+
"grad_norm": 0.4586232602596283,
|
270 |
+
"learning_rate": 0.0002918251435864709,
|
271 |
+
"loss": 2.3499,
|
272 |
+
"step": 35000
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 2.5407579928011854,
|
276 |
+
"grad_norm": 0.4139242172241211,
|
277 |
+
"learning_rate": 0.0002890928981675631,
|
278 |
+
"loss": 2.3302,
|
279 |
+
"step": 36000
|
280 |
+
},
|
281 |
+
{
|
282 |
+
"epoch": 2.6113346037123297,
|
283 |
+
"grad_norm": 0.41075077652931213,
|
284 |
+
"learning_rate": 0.000286357917768256,
|
285 |
+
"loss": 2.3236,
|
286 |
+
"step": 37000
|
287 |
+
},
|
288 |
+
{
|
289 |
+
"epoch": 2.681911214623474,
|
290 |
+
"grad_norm": 0.4315313994884491,
|
291 |
+
"learning_rate": 0.00028362567234934815,
|
292 |
+
"loss": 2.3239,
|
293 |
+
"step": 38000
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 2.752487825534618,
|
297 |
+
"grad_norm": 0.39671897888183594,
|
298 |
+
"learning_rate": 0.000280890691950041,
|
299 |
+
"loss": 2.3087,
|
300 |
+
"step": 39000
|
301 |
+
},
|
302 |
+
{
|
303 |
+
"epoch": 2.8230644364457618,
|
304 |
+
"grad_norm": 0.4033380150794983,
|
305 |
+
"learning_rate": 0.00027815844653113317,
|
306 |
+
"loss": 2.2952,
|
307 |
+
"step": 40000
|
308 |
+
},
|
309 |
+
{
|
310 |
+
"epoch": 2.893641047356906,
|
311 |
+
"grad_norm": 0.36479175090789795,
|
312 |
+
"learning_rate": 0.000275423466131826,
|
313 |
+
"loss": 2.3049,
|
314 |
+
"step": 41000
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.96421765826805,
|
318 |
+
"grad_norm": 0.3533291220664978,
|
319 |
+
"learning_rate": 0.00027268848573251887,
|
320 |
+
"loss": 2.286,
|
321 |
+
"step": 42000
|
322 |
+
},
|
323 |
+
{
|
324 |
+
"epoch": 3.0,
|
325 |
+
"eval_accuracy": 0.5122970740171453,
|
326 |
+
"eval_loss": 2.5088469982147217,
|
327 |
+
"eval_runtime": 124.5088,
|
328 |
+
"eval_samples_per_second": 377.09,
|
329 |
+
"eval_steps_per_second": 5.895,
|
330 |
+
"step": 42507
|
331 |
+
},
|
332 |
+
{
|
333 |
+
"epoch": 3.034794269179194,
|
334 |
+
"grad_norm": 0.37031927704811096,
|
335 |
+
"learning_rate": 0.00026995350533321177,
|
336 |
+
"loss": 2.2529,
|
337 |
+
"step": 43000
|
338 |
+
},
|
339 |
+
{
|
340 |
+
"epoch": 3.105370880090338,
|
341 |
+
"grad_norm": 0.39132222533226013,
|
342 |
+
"learning_rate": 0.0002672185249339046,
|
343 |
+
"loss": 2.2252,
|
344 |
+
"step": 44000
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 3.175947491001482,
|
348 |
+
"grad_norm": 0.36771416664123535,
|
349 |
+
"learning_rate": 0.0002644835445345975,
|
350 |
+
"loss": 2.2389,
|
351 |
+
"step": 45000
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"epoch": 3.2465241019126263,
|
355 |
+
"grad_norm": 0.40541785955429077,
|
356 |
+
"learning_rate": 0.00026175129911568964,
|
357 |
+
"loss": 2.2224,
|
358 |
+
"step": 46000
|
359 |
+
},
|
360 |
+
{
|
361 |
+
"epoch": 3.31710071282377,
|
362 |
+
"grad_norm": 0.4325103461742401,
|
363 |
+
"learning_rate": 0.0002590190536967818,
|
364 |
+
"loss": 2.2286,
|
365 |
+
"step": 47000
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 3.3876773237349145,
|
369 |
+
"grad_norm": 0.3990887403488159,
|
370 |
+
"learning_rate": 0.00025628407329747466,
|
371 |
+
"loss": 2.216,
|
372 |
+
"step": 48000
|
373 |
+
},
|
374 |
+
{
|
375 |
+
"epoch": 3.4582539346460583,
|
376 |
+
"grad_norm": 0.36805373430252075,
|
377 |
+
"learning_rate": 0.00025354909289816756,
|
378 |
+
"loss": 2.2209,
|
379 |
+
"step": 49000
|
380 |
+
},
|
381 |
+
{
|
382 |
+
"epoch": 3.528830545557202,
|
383 |
+
"grad_norm": 0.357721209526062,
|
384 |
+
"learning_rate": 0.0002508141124988604,
|
385 |
+
"loss": 2.2218,
|
386 |
+
"step": 50000
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 3.5994071564683465,
|
390 |
+
"grad_norm": 0.3424566984176636,
|
391 |
+
"learning_rate": 0.00024808460206035186,
|
392 |
+
"loss": 2.2159,
|
393 |
+
"step": 51000
|
394 |
+
},
|
395 |
+
{
|
396 |
+
"epoch": 3.6699837673794904,
|
397 |
+
"grad_norm": 0.35358792543411255,
|
398 |
+
"learning_rate": 0.00024534962166104476,
|
399 |
+
"loss": 2.2085,
|
400 |
+
"step": 52000
|
401 |
+
},
|
402 |
+
{
|
403 |
+
"epoch": 3.7405603782906347,
|
404 |
+
"grad_norm": 0.369150310754776,
|
405 |
+
"learning_rate": 0.0002426146412617376,
|
406 |
+
"loss": 2.2033,
|
407 |
+
"step": 53000
|
408 |
+
},
|
409 |
+
{
|
410 |
+
"epoch": 3.8111369892017786,
|
411 |
+
"grad_norm": 0.3341532051563263,
|
412 |
+
"learning_rate": 0.00023987966086243048,
|
413 |
+
"loss": 2.2017,
|
414 |
+
"step": 54000
|
415 |
+
},
|
416 |
+
{
|
417 |
+
"epoch": 3.8817136001129224,
|
418 |
+
"grad_norm": 0.40089789032936096,
|
419 |
+
"learning_rate": 0.00023714741544352263,
|
420 |
+
"loss": 2.2015,
|
421 |
+
"step": 55000
|
422 |
+
},
|
423 |
+
{
|
424 |
+
"epoch": 3.9522902110240667,
|
425 |
+
"grad_norm": 0.35854268074035645,
|
426 |
+
"learning_rate": 0.0002344124350442155,
|
427 |
+
"loss": 2.2124,
|
428 |
+
"step": 56000
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 4.0,
|
432 |
+
"eval_accuracy": 0.5203132962446899,
|
433 |
+
"eval_loss": 2.444835662841797,
|
434 |
+
"eval_runtime": 124.4717,
|
435 |
+
"eval_samples_per_second": 377.202,
|
436 |
+
"eval_steps_per_second": 5.897,
|
437 |
+
"step": 56676
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 4.022866821935211,
|
441 |
+
"grad_norm": 0.3867768347263336,
|
442 |
+
"learning_rate": 0.00023167745464490835,
|
443 |
+
"loss": 2.1839,
|
444 |
+
"step": 57000
|
445 |
+
},
|
446 |
+
{
|
447 |
+
"epoch": 4.093443432846355,
|
448 |
+
"grad_norm": 0.3627295196056366,
|
449 |
+
"learning_rate": 0.00022894520922600055,
|
450 |
+
"loss": 2.1384,
|
451 |
+
"step": 58000
|
452 |
+
},
|
453 |
+
{
|
454 |
+
"epoch": 4.164020043757499,
|
455 |
+
"grad_norm": 0.35046979784965515,
|
456 |
+
"learning_rate": 0.00022621022882669337,
|
457 |
+
"loss": 2.1439,
|
458 |
+
"step": 59000
|
459 |
+
},
|
460 |
+
{
|
461 |
+
"epoch": 4.234596654668643,
|
462 |
+
"grad_norm": 0.3317088186740875,
|
463 |
+
"learning_rate": 0.00022347524842738624,
|
464 |
+
"loss": 2.1446,
|
465 |
+
"step": 60000
|
466 |
+
},
|
467 |
+
{
|
468 |
+
"epoch": 4.3051732655797865,
|
469 |
+
"grad_norm": 0.37563446164131165,
|
470 |
+
"learning_rate": 0.00022074300300847845,
|
471 |
+
"loss": 2.1408,
|
472 |
+
"step": 61000
|
473 |
+
},
|
474 |
+
{
|
475 |
+
"epoch": 4.375749876490931,
|
476 |
+
"grad_norm": 0.36360374093055725,
|
477 |
+
"learning_rate": 0.00021800802260917127,
|
478 |
+
"loss": 2.1447,
|
479 |
+
"step": 62000
|
480 |
+
},
|
481 |
+
{
|
482 |
+
"epoch": 4.446326487402075,
|
483 |
+
"grad_norm": 0.35474300384521484,
|
484 |
+
"learning_rate": 0.00021527577719026347,
|
485 |
+
"loss": 2.1478,
|
486 |
+
"step": 63000
|
487 |
+
},
|
488 |
+
{
|
489 |
+
"epoch": 4.516903098313219,
|
490 |
+
"grad_norm": 0.38771218061447144,
|
491 |
+
"learning_rate": 0.00021254079679095632,
|
492 |
+
"loss": 2.136,
|
493 |
+
"step": 64000
|
494 |
+
},
|
495 |
+
{
|
496 |
+
"epoch": 4.587479709224363,
|
497 |
+
"grad_norm": 0.3860458433628082,
|
498 |
+
"learning_rate": 0.0002098085513720485,
|
499 |
+
"loss": 2.1379,
|
500 |
+
"step": 65000
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 4.658056320135507,
|
504 |
+
"grad_norm": 0.3763484060764313,
|
505 |
+
"learning_rate": 0.00020707357097274134,
|
506 |
+
"loss": 2.1344,
|
507 |
+
"step": 66000
|
508 |
+
},
|
509 |
+
{
|
510 |
+
"epoch": 4.7286329310466515,
|
511 |
+
"grad_norm": 0.37907466292381287,
|
512 |
+
"learning_rate": 0.0002043413255538335,
|
513 |
+
"loss": 2.1426,
|
514 |
+
"step": 67000
|
515 |
+
},
|
516 |
+
{
|
517 |
+
"epoch": 4.799209541957795,
|
518 |
+
"grad_norm": 0.34690865874290466,
|
519 |
+
"learning_rate": 0.00020160634515452636,
|
520 |
+
"loss": 2.1208,
|
521 |
+
"step": 68000
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 4.869786152868939,
|
525 |
+
"grad_norm": 0.36183568835258484,
|
526 |
+
"learning_rate": 0.00019887409973561856,
|
527 |
+
"loss": 2.1242,
|
528 |
+
"step": 69000
|
529 |
+
},
|
530 |
+
{
|
531 |
+
"epoch": 4.940362763780083,
|
532 |
+
"grad_norm": 0.35947293043136597,
|
533 |
+
"learning_rate": 0.00019613911933631138,
|
534 |
+
"loss": 2.1307,
|
535 |
+
"step": 70000
|
536 |
+
},
|
537 |
+
{
|
538 |
+
"epoch": 5.0,
|
539 |
+
"eval_accuracy": 0.5251367702083976,
|
540 |
+
"eval_loss": 2.4099230766296387,
|
541 |
+
"eval_runtime": 124.5048,
|
542 |
+
"eval_samples_per_second": 377.102,
|
543 |
+
"eval_steps_per_second": 5.895,
|
544 |
+
"step": 70845
|
545 |
+
},
|
546 |
+
{
|
547 |
+
"epoch": 5.010939374691227,
|
548 |
+
"grad_norm": 0.3575204312801361,
|
549 |
+
"learning_rate": 0.00019340687391740358,
|
550 |
+
"loss": 2.1265,
|
551 |
+
"step": 71000
|
552 |
+
},
|
553 |
+
{
|
554 |
+
"epoch": 5.081515985602372,
|
555 |
+
"grad_norm": 0.38784071803092957,
|
556 |
+
"learning_rate": 0.00019067189351809646,
|
557 |
+
"loss": 2.064,
|
558 |
+
"step": 72000
|
559 |
+
},
|
560 |
+
{
|
561 |
+
"epoch": 5.152092596513516,
|
562 |
+
"grad_norm": 0.34589263796806335,
|
563 |
+
"learning_rate": 0.0001879396480991886,
|
564 |
+
"loss": 2.0754,
|
565 |
+
"step": 73000
|
566 |
+
},
|
567 |
+
{
|
568 |
+
"epoch": 5.2226692074246595,
|
569 |
+
"grad_norm": 0.3403594195842743,
|
570 |
+
"learning_rate": 0.00018520466769988148,
|
571 |
+
"loss": 2.073,
|
572 |
+
"step": 74000
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 5.293245818335803,
|
576 |
+
"grad_norm": 0.38706710934638977,
|
577 |
+
"learning_rate": 0.00018246968730057433,
|
578 |
+
"loss": 2.0812,
|
579 |
+
"step": 75000
|
580 |
+
},
|
581 |
+
{
|
582 |
+
"epoch": 5.363822429246947,
|
583 |
+
"grad_norm": 0.39309191703796387,
|
584 |
+
"learning_rate": 0.0001797347069012672,
|
585 |
+
"loss": 2.0915,
|
586 |
+
"step": 76000
|
587 |
+
},
|
588 |
+
{
|
589 |
+
"epoch": 5.434399040158092,
|
590 |
+
"grad_norm": 0.37432822585105896,
|
591 |
+
"learning_rate": 0.00017700246148235935,
|
592 |
+
"loss": 2.0755,
|
593 |
+
"step": 77000
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 5.504975651069236,
|
597 |
+
"grad_norm": 0.3538018465042114,
|
598 |
+
"learning_rate": 0.00017426748108305222,
|
599 |
+
"loss": 2.0772,
|
600 |
+
"step": 78000
|
601 |
+
},
|
602 |
+
{
|
603 |
+
"epoch": 5.57555226198038,
|
604 |
+
"grad_norm": 0.3601301610469818,
|
605 |
+
"learning_rate": 0.00017153523566414437,
|
606 |
+
"loss": 2.085,
|
607 |
+
"step": 79000
|
608 |
+
},
|
609 |
+
{
|
610 |
+
"epoch": 5.6461288728915235,
|
611 |
+
"grad_norm": 0.3469400703907013,
|
612 |
+
"learning_rate": 0.00016880299024523657,
|
613 |
+
"loss": 2.078,
|
614 |
+
"step": 80000
|
615 |
+
},
|
616 |
+
{
|
617 |
+
"epoch": 5.716705483802667,
|
618 |
+
"grad_norm": 0.3623177111148834,
|
619 |
+
"learning_rate": 0.00016607074482632872,
|
620 |
+
"loss": 2.077,
|
621 |
+
"step": 81000
|
622 |
+
},
|
623 |
+
{
|
624 |
+
"epoch": 5.787282094713812,
|
625 |
+
"grad_norm": 0.3382331132888794,
|
626 |
+
"learning_rate": 0.0001633357644270216,
|
627 |
+
"loss": 2.0745,
|
628 |
+
"step": 82000
|
629 |
+
},
|
630 |
+
{
|
631 |
+
"epoch": 5.857858705624956,
|
632 |
+
"grad_norm": 0.3787217140197754,
|
633 |
+
"learning_rate": 0.00016060078402771447,
|
634 |
+
"loss": 2.0768,
|
635 |
+
"step": 83000
|
636 |
+
},
|
637 |
+
{
|
638 |
+
"epoch": 5.9284353165361,
|
639 |
+
"grad_norm": 0.36773914098739624,
|
640 |
+
"learning_rate": 0.00015786580362840732,
|
641 |
+
"loss": 2.0756,
|
642 |
+
"step": 84000
|
643 |
+
},
|
644 |
+
{
|
645 |
+
"epoch": 5.999011927447244,
|
646 |
+
"grad_norm": 0.39036858081817627,
|
647 |
+
"learning_rate": 0.00015513082322910016,
|
648 |
+
"loss": 2.0706,
|
649 |
+
"step": 85000
|
650 |
+
},
|
651 |
+
{
|
652 |
+
"epoch": 6.0,
|
653 |
+
"eval_accuracy": 0.5280841264512295,
|
654 |
+
"eval_loss": 2.388700246810913,
|
655 |
+
"eval_runtime": 125.3022,
|
656 |
+
"eval_samples_per_second": 374.702,
|
657 |
+
"eval_steps_per_second": 5.858,
|
658 |
+
"step": 85014
|
659 |
+
},
|
660 |
+
{
|
661 |
+
"epoch": 6.069588538358388,
|
662 |
+
"grad_norm": 0.3616304099559784,
|
663 |
+
"learning_rate": 0.00015239584282979304,
|
664 |
+
"loss": 2.0123,
|
665 |
+
"step": 86000
|
666 |
+
},
|
667 |
+
{
|
668 |
+
"epoch": 6.140165149269532,
|
669 |
+
"grad_norm": 0.37002402544021606,
|
670 |
+
"learning_rate": 0.0001496635974108852,
|
671 |
+
"loss": 2.0153,
|
672 |
+
"step": 87000
|
673 |
+
},
|
674 |
+
{
|
675 |
+
"epoch": 6.210741760180676,
|
676 |
+
"grad_norm": 0.3603184223175049,
|
677 |
+
"learning_rate": 0.0001469286170115781,
|
678 |
+
"loss": 2.0199,
|
679 |
+
"step": 88000
|
680 |
+
},
|
681 |
+
{
|
682 |
+
"epoch": 6.28131837109182,
|
683 |
+
"grad_norm": 0.37550196051597595,
|
684 |
+
"learning_rate": 0.00014419637159267024,
|
685 |
+
"loss": 2.0138,
|
686 |
+
"step": 89000
|
687 |
+
},
|
688 |
+
{
|
689 |
+
"epoch": 6.351894982002964,
|
690 |
+
"grad_norm": 0.3768686056137085,
|
691 |
+
"learning_rate": 0.0001414641261737624,
|
692 |
+
"loss": 2.0314,
|
693 |
+
"step": 90000
|
694 |
+
},
|
695 |
+
{
|
696 |
+
"epoch": 6.422471592914108,
|
697 |
+
"grad_norm": 0.3591078221797943,
|
698 |
+
"learning_rate": 0.00013872914577445526,
|
699 |
+
"loss": 2.0226,
|
700 |
+
"step": 91000
|
701 |
+
},
|
702 |
+
{
|
703 |
+
"epoch": 6.493048203825253,
|
704 |
+
"grad_norm": 0.3905663788318634,
|
705 |
+
"learning_rate": 0.00013599416537514813,
|
706 |
+
"loss": 2.0257,
|
707 |
+
"step": 92000
|
708 |
+
},
|
709 |
+
{
|
710 |
+
"epoch": 6.5636248147363965,
|
711 |
+
"grad_norm": 0.39147230982780457,
|
712 |
+
"learning_rate": 0.00013325918497584098,
|
713 |
+
"loss": 2.0311,
|
714 |
+
"step": 93000
|
715 |
+
},
|
716 |
+
{
|
717 |
+
"epoch": 6.63420142564754,
|
718 |
+
"grad_norm": 0.40250155329704285,
|
719 |
+
"learning_rate": 0.00013052420457653385,
|
720 |
+
"loss": 2.0301,
|
721 |
+
"step": 94000
|
722 |
+
},
|
723 |
+
{
|
724 |
+
"epoch": 6.704778036558684,
|
725 |
+
"grad_norm": 0.3860897123813629,
|
726 |
+
"learning_rate": 0.00012779195915762603,
|
727 |
+
"loss": 2.0312,
|
728 |
+
"step": 95000
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 6.775354647469829,
|
732 |
+
"grad_norm": 0.3707718253135681,
|
733 |
+
"learning_rate": 0.0001250597137387182,
|
734 |
+
"loss": 2.0282,
|
735 |
+
"step": 96000
|
736 |
+
},
|
737 |
+
{
|
738 |
+
"epoch": 6.845931258380973,
|
739 |
+
"grad_norm": 0.3690090775489807,
|
740 |
+
"learning_rate": 0.00012232473333941105,
|
741 |
+
"loss": 2.0255,
|
742 |
+
"step": 97000
|
743 |
+
},
|
744 |
+
{
|
745 |
+
"epoch": 6.916507869292117,
|
746 |
+
"grad_norm": 0.4174107313156128,
|
747 |
+
"learning_rate": 0.00011958975294010391,
|
748 |
+
"loss": 2.0276,
|
749 |
+
"step": 98000
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 6.987084480203261,
|
753 |
+
"grad_norm": 0.3740842938423157,
|
754 |
+
"learning_rate": 0.00011685750752119609,
|
755 |
+
"loss": 2.0233,
|
756 |
+
"step": 99000
|
757 |
+
},
|
758 |
+
{
|
759 |
+
"epoch": 7.0,
|
760 |
+
"eval_accuracy": 0.5303788443403243,
|
761 |
+
"eval_loss": 2.3779311180114746,
|
762 |
+
"eval_runtime": 124.6993,
|
763 |
+
"eval_samples_per_second": 376.514,
|
764 |
+
"eval_steps_per_second": 5.886,
|
765 |
+
"step": 99183
|
766 |
+
},
|
767 |
+
{
|
768 |
+
"epoch": 7.057661091114404,
|
769 |
+
"grad_norm": 0.37689074873924255,
|
770 |
+
"learning_rate": 0.00011412252712188895,
|
771 |
+
"loss": 1.9805,
|
772 |
+
"step": 100000
|
773 |
+
},
|
774 |
+
{
|
775 |
+
"epoch": 7.128237702025549,
|
776 |
+
"grad_norm": 0.4047756493091583,
|
777 |
+
"learning_rate": 0.00011139028170298112,
|
778 |
+
"loss": 1.9746,
|
779 |
+
"step": 101000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 7.198814312936693,
|
783 |
+
"grad_norm": 0.3889460563659668,
|
784 |
+
"learning_rate": 0.00010865530130367397,
|
785 |
+
"loss": 1.9659,
|
786 |
+
"step": 102000
|
787 |
+
},
|
788 |
+
{
|
789 |
+
"epoch": 7.269390923847837,
|
790 |
+
"grad_norm": 0.4061487019062042,
|
791 |
+
"learning_rate": 0.00010592032090436684,
|
792 |
+
"loss": 1.9708,
|
793 |
+
"step": 103000
|
794 |
+
},
|
795 |
+
{
|
796 |
+
"epoch": 7.339967534758981,
|
797 |
+
"grad_norm": 0.4057160019874573,
|
798 |
+
"learning_rate": 0.0001031853405050597,
|
799 |
+
"loss": 1.9772,
|
800 |
+
"step": 104000
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 7.410544145670125,
|
804 |
+
"grad_norm": 0.40489134192466736,
|
805 |
+
"learning_rate": 0.00010045309508615188,
|
806 |
+
"loss": 1.9736,
|
807 |
+
"step": 105000
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 7.481120756581269,
|
811 |
+
"grad_norm": 0.4148654043674469,
|
812 |
+
"learning_rate": 9.771811468684473e-05,
|
813 |
+
"loss": 1.9796,
|
814 |
+
"step": 106000
|
815 |
+
},
|
816 |
+
{
|
817 |
+
"epoch": 7.551697367492413,
|
818 |
+
"grad_norm": 0.41483381390571594,
|
819 |
+
"learning_rate": 9.49831342875376e-05,
|
820 |
+
"loss": 1.9804,
|
821 |
+
"step": 107000
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 7.622273978403557,
|
825 |
+
"grad_norm": 0.43718773126602173,
|
826 |
+
"learning_rate": 9.225088886862978e-05,
|
827 |
+
"loss": 1.9712,
|
828 |
+
"step": 108000
|
829 |
+
},
|
830 |
+
{
|
831 |
+
"epoch": 7.692850589314701,
|
832 |
+
"grad_norm": 0.40646329522132874,
|
833 |
+
"learning_rate": 8.951590846932262e-05,
|
834 |
+
"loss": 1.9822,
|
835 |
+
"step": 109000
|
836 |
+
},
|
837 |
+
{
|
838 |
+
"epoch": 7.763427200225845,
|
839 |
+
"grad_norm": 0.44571158289909363,
|
840 |
+
"learning_rate": 8.67836630504148e-05,
|
841 |
+
"loss": 1.9832,
|
842 |
+
"step": 110000
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 7.83400381113699,
|
846 |
+
"grad_norm": 0.41726765036582947,
|
847 |
+
"learning_rate": 8.404868265110766e-05,
|
848 |
+
"loss": 1.9747,
|
849 |
+
"step": 111000
|
850 |
+
},
|
851 |
+
{
|
852 |
+
"epoch": 7.9045804220481335,
|
853 |
+
"grad_norm": 0.39210569858551025,
|
854 |
+
"learning_rate": 8.131370225180053e-05,
|
855 |
+
"loss": 1.9929,
|
856 |
+
"step": 112000
|
857 |
+
},
|
858 |
+
{
|
859 |
+
"epoch": 7.975157032959277,
|
860 |
+
"grad_norm": 0.37121346592903137,
|
861 |
+
"learning_rate": 7.858145683289268e-05,
|
862 |
+
"loss": 1.9727,
|
863 |
+
"step": 113000
|
864 |
+
},
|
865 |
+
{
|
866 |
+
"epoch": 8.0,
|
867 |
+
"eval_accuracy": 0.5315104156523147,
|
868 |
+
"eval_loss": 2.3731467723846436,
|
869 |
+
"eval_runtime": 124.474,
|
870 |
+
"eval_samples_per_second": 377.195,
|
871 |
+
"eval_steps_per_second": 5.897,
|
872 |
+
"step": 113352
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 8.045733643870422,
|
876 |
+
"grad_norm": 0.3864983916282654,
|
877 |
+
"learning_rate": 7.584647643358555e-05,
|
878 |
+
"loss": 1.9365,
|
879 |
+
"step": 114000
|
880 |
+
},
|
881 |
+
{
|
882 |
+
"epoch": 8.116310254781565,
|
883 |
+
"grad_norm": 0.4133272171020508,
|
884 |
+
"learning_rate": 7.311423101467773e-05,
|
885 |
+
"loss": 1.9174,
|
886 |
+
"step": 115000
|
887 |
+
},
|
888 |
+
{
|
889 |
+
"epoch": 8.18688686569271,
|
890 |
+
"grad_norm": 0.4471355676651001,
|
891 |
+
"learning_rate": 7.037925061537059e-05,
|
892 |
+
"loss": 1.9327,
|
893 |
+
"step": 116000
|
894 |
+
},
|
895 |
+
{
|
896 |
+
"epoch": 8.257463476603853,
|
897 |
+
"grad_norm": 0.42897623777389526,
|
898 |
+
"learning_rate": 6.764700519646275e-05,
|
899 |
+
"loss": 1.9288,
|
900 |
+
"step": 117000
|
901 |
+
},
|
902 |
+
{
|
903 |
+
"epoch": 8.328040087514998,
|
904 |
+
"grad_norm": 0.43864506483078003,
|
905 |
+
"learning_rate": 6.491202479715561e-05,
|
906 |
+
"loss": 1.9359,
|
907 |
+
"step": 118000
|
908 |
+
},
|
909 |
+
{
|
910 |
+
"epoch": 8.398616698426142,
|
911 |
+
"grad_norm": 0.46767184138298035,
|
912 |
+
"learning_rate": 6.217977937824779e-05,
|
913 |
+
"loss": 1.9399,
|
914 |
+
"step": 119000
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 8.469193309337285,
|
918 |
+
"grad_norm": 0.4159405827522278,
|
919 |
+
"learning_rate": 5.944479897894064e-05,
|
920 |
+
"loss": 1.9318,
|
921 |
+
"step": 120000
|
922 |
+
},
|
923 |
+
{
|
924 |
+
"epoch": 8.53976992024843,
|
925 |
+
"grad_norm": 0.4233142137527466,
|
926 |
+
"learning_rate": 5.671255356003281e-05,
|
927 |
+
"loss": 1.9316,
|
928 |
+
"step": 121000
|
929 |
+
},
|
930 |
+
{
|
931 |
+
"epoch": 8.610346531159573,
|
932 |
+
"grad_norm": 0.4398587942123413,
|
933 |
+
"learning_rate": 5.3980308141124983e-05,
|
934 |
+
"loss": 1.9349,
|
935 |
+
"step": 122000
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 8.680923142070718,
|
939 |
+
"grad_norm": 0.424790620803833,
|
940 |
+
"learning_rate": 5.124532774181785e-05,
|
941 |
+
"loss": 1.9382,
|
942 |
+
"step": 123000
|
943 |
+
},
|
944 |
+
{
|
945 |
+
"epoch": 8.751499752981863,
|
946 |
+
"grad_norm": 0.4141928553581238,
|
947 |
+
"learning_rate": 4.851034734251071e-05,
|
948 |
+
"loss": 1.9314,
|
949 |
+
"step": 124000
|
950 |
+
},
|
951 |
+
{
|
952 |
+
"epoch": 8.822076363893006,
|
953 |
+
"grad_norm": 0.45448678731918335,
|
954 |
+
"learning_rate": 4.577536694320357e-05,
|
955 |
+
"loss": 1.9374,
|
956 |
+
"step": 125000
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 8.89265297480415,
|
960 |
+
"grad_norm": 0.4196777939796448,
|
961 |
+
"learning_rate": 4.304312152429574e-05,
|
962 |
+
"loss": 1.9413,
|
963 |
+
"step": 126000
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 8.963229585715293,
|
967 |
+
"grad_norm": 0.3975803256034851,
|
968 |
+
"learning_rate": 4.03081411249886e-05,
|
969 |
+
"loss": 1.9311,
|
970 |
+
"step": 127000
|
971 |
+
},
|
972 |
+
{
|
973 |
+
"epoch": 9.0,
|
974 |
+
"eval_accuracy": 0.532392202583586,
|
975 |
+
"eval_loss": 2.3728187084198,
|
976 |
+
"eval_runtime": 124.6433,
|
977 |
+
"eval_samples_per_second": 376.683,
|
978 |
+
"eval_steps_per_second": 5.889,
|
979 |
+
"step": 127521
|
980 |
+
},
|
981 |
+
{
|
982 |
+
"epoch": 9.033806196626438,
|
983 |
+
"grad_norm": 0.42701438069343567,
|
984 |
+
"learning_rate": 3.757589570608077e-05,
|
985 |
+
"loss": 1.9097,
|
986 |
+
"step": 128000
|
987 |
+
},
|
988 |
+
{
|
989 |
+
"epoch": 9.104382807537583,
|
990 |
+
"grad_norm": 0.44095513224601746,
|
991 |
+
"learning_rate": 3.484091530677363e-05,
|
992 |
+
"loss": 1.8841,
|
993 |
+
"step": 129000
|
994 |
+
},
|
995 |
+
{
|
996 |
+
"epoch": 9.174959418448726,
|
997 |
+
"grad_norm": 0.4554646909236908,
|
998 |
+
"learning_rate": 3.2108669887865805e-05,
|
999 |
+
"loss": 1.898,
|
1000 |
+
"step": 130000
|
1001 |
+
},
|
1002 |
+
{
|
1003 |
+
"epoch": 9.24553602935987,
|
1004 |
+
"grad_norm": 0.4306824803352356,
|
1005 |
+
"learning_rate": 2.937642446895797e-05,
|
1006 |
+
"loss": 1.8921,
|
1007 |
+
"step": 131000
|
1008 |
+
},
|
1009 |
+
{
|
1010 |
+
"epoch": 9.316112640271013,
|
1011 |
+
"grad_norm": 0.4496276080608368,
|
1012 |
+
"learning_rate": 2.664144406965083e-05,
|
1013 |
+
"loss": 1.8883,
|
1014 |
+
"step": 132000
|
1015 |
+
},
|
1016 |
+
{
|
1017 |
+
"epoch": 9.386689251182158,
|
1018 |
+
"grad_norm": 0.4314197301864624,
|
1019 |
+
"learning_rate": 2.3906463670343695e-05,
|
1020 |
+
"loss": 1.8972,
|
1021 |
+
"step": 133000
|
1022 |
+
},
|
1023 |
+
{
|
1024 |
+
"epoch": 9.457265862093303,
|
1025 |
+
"grad_norm": 0.4485911428928375,
|
1026 |
+
"learning_rate": 2.1171483271036556e-05,
|
1027 |
+
"loss": 1.8908,
|
1028 |
+
"step": 134000
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 9.527842473004446,
|
1032 |
+
"grad_norm": 0.4166420102119446,
|
1033 |
+
"learning_rate": 1.8439237852128724e-05,
|
1034 |
+
"loss": 1.8924,
|
1035 |
+
"step": 135000
|
1036 |
+
},
|
1037 |
+
{
|
1038 |
+
"epoch": 9.59841908391559,
|
1039 |
+
"grad_norm": 0.4447433650493622,
|
1040 |
+
"learning_rate": 1.5704257452821588e-05,
|
1041 |
+
"loss": 1.892,
|
1042 |
+
"step": 136000
|
1043 |
+
},
|
1044 |
+
{
|
1045 |
+
"epoch": 9.668995694826734,
|
1046 |
+
"grad_norm": 0.42986035346984863,
|
1047 |
+
"learning_rate": 1.2972012033913756e-05,
|
1048 |
+
"loss": 1.8992,
|
1049 |
+
"step": 137000
|
1050 |
+
},
|
1051 |
+
{
|
1052 |
+
"epoch": 9.739572305737878,
|
1053 |
+
"grad_norm": 0.45574498176574707,
|
1054 |
+
"learning_rate": 1.0237031634606618e-05,
|
1055 |
+
"loss": 1.8972,
|
1056 |
+
"step": 138000
|
1057 |
+
},
|
1058 |
+
{
|
1059 |
+
"epoch": 9.810148916649023,
|
1060 |
+
"grad_norm": 0.4515238404273987,
|
1061 |
+
"learning_rate": 7.5020512352994795e-06,
|
1062 |
+
"loss": 1.8952,
|
1063 |
+
"step": 139000
|
1064 |
+
},
|
1065 |
+
{
|
1066 |
+
"epoch": 9.880725527560166,
|
1067 |
+
"grad_norm": 0.43244990706443787,
|
1068 |
+
"learning_rate": 4.769805816391649e-06,
|
1069 |
+
"loss": 1.8946,
|
1070 |
+
"step": 140000
|
1071 |
+
},
|
1072 |
+
{
|
1073 |
+
"epoch": 9.951302138471311,
|
1074 |
+
"grad_norm": 0.44382914900779724,
|
1075 |
+
"learning_rate": 2.0348254170845108e-06,
|
1076 |
+
"loss": 1.8943,
|
1077 |
+
"step": 141000
|
1078 |
+
},
|
1079 |
+
{
|
1080 |
+
"epoch": 10.0,
|
1081 |
+
"eval_accuracy": 0.5327396962492645,
|
1082 |
+
"eval_loss": 2.377573251724243,
|
1083 |
+
"eval_runtime": 124.448,
|
1084 |
+
"eval_samples_per_second": 377.274,
|
1085 |
+
"eval_steps_per_second": 5.898,
|
1086 |
+
"step": 141690
|
1087 |
+
},
|
1088 |
+
{
|
1089 |
+
"epoch": 10.0,
|
1090 |
+
"step": 141690,
|
1091 |
+
"total_flos": 5.9695438005504e+17,
|
1092 |
+
"train_loss": 2.237272139724321,
|
1093 |
+
"train_runtime": 31036.7572,
|
1094 |
+
"train_samples_per_second": 146.079,
|
1095 |
+
"train_steps_per_second": 4.565
|
1096 |
+
}
|
1097 |
+
],
|
1098 |
+
"logging_steps": 1000,
|
1099 |
+
"max_steps": 141690,
|
1100 |
+
"num_input_tokens_seen": 0,
|
1101 |
+
"num_train_epochs": 10,
|
1102 |
+
"save_steps": 5000,
|
1103 |
+
"stateful_callbacks": {
|
1104 |
+
"TrainerControl": {
|
1105 |
+
"args": {
|
1106 |
+
"should_epoch_stop": false,
|
1107 |
+
"should_evaluate": false,
|
1108 |
+
"should_log": false,
|
1109 |
+
"should_save": true,
|
1110 |
+
"should_training_stop": true
|
1111 |
+
},
|
1112 |
+
"attributes": {}
|
1113 |
+
}
|
1114 |
+
},
|
1115 |
+
"total_flos": 5.9695438005504e+17,
|
1116 |
+
"train_batch_size": 32,
|
1117 |
+
"trial_name": null,
|
1118 |
+
"trial_params": null
|
1119 |
+
}
|