nkuAlexLee
commited on
Commit
•
7456fc1
1
Parent(s):
c0207db
Upload 13 files
Browse files- config.json +1 -1
- optimizer.pt +2 -2
- pytorch_model.bin +1 -1
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +66 -0
- training_args.bin +2 -2
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_dropout": 0.0,
|
4 |
"activation_function": "swish",
|
5 |
"add_bias_logits": false,
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "DDDSSS/translation_en-zh",
|
3 |
"activation_dropout": 0.0,
|
4 |
"activation_function": "swish",
|
5 |
"add_bias_logits": false,
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f25bf36a56ac69410a3763cc4eace80e0aa12b75a85c46f41dc6b3f104a6538e
|
3 |
+
size 619500549
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 310022533
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2bbc0a089851cae6fbc14a78afae6fb0fa036c80699bf4fc9edd49b24c87843f
|
3 |
size 310022533
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4369e459495ac86f5932270534e063e608774e09de46bc38426a0e61c4a3153b
|
3 |
+
size 14575
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:71ce1e82580d06836277ac609a46ef220afb039c95e4479b9265ae1045ab90e0
|
3 |
+
size 627
|
trainer_state.json
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 4.0,
|
5 |
+
"global_step": 572,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 1.0,
|
12 |
+
"eval_bleu": 7.7545,
|
13 |
+
"eval_gen_len": 14.8741,
|
14 |
+
"eval_jit_compilation_time": 11.4032,
|
15 |
+
"eval_loss": 2.9266164302825928,
|
16 |
+
"eval_runtime": 12.5758,
|
17 |
+
"eval_samples_per_second": 11.371,
|
18 |
+
"eval_steps_per_second": 2.863,
|
19 |
+
"step": 143
|
20 |
+
},
|
21 |
+
{
|
22 |
+
"epoch": 2.0,
|
23 |
+
"eval_bleu": 10.4293,
|
24 |
+
"eval_gen_len": 14.0629,
|
25 |
+
"eval_jit_compilation_time": 12.1037,
|
26 |
+
"eval_loss": 2.850106716156006,
|
27 |
+
"eval_runtime": 11.8772,
|
28 |
+
"eval_samples_per_second": 12.04,
|
29 |
+
"eval_steps_per_second": 3.031,
|
30 |
+
"step": 286
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 3.0,
|
34 |
+
"eval_bleu": 11.0503,
|
35 |
+
"eval_gen_len": 14.5664,
|
36 |
+
"eval_jit_compilation_time": 12.7826,
|
37 |
+
"eval_loss": 2.875509262084961,
|
38 |
+
"eval_runtime": 12.4472,
|
39 |
+
"eval_samples_per_second": 11.489,
|
40 |
+
"eval_steps_per_second": 2.892,
|
41 |
+
"step": 429
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 3.5,
|
45 |
+
"learning_rate": 2.5874125874125877e-05,
|
46 |
+
"loss": 2.0491,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 4.0,
|
51 |
+
"eval_bleu": 12.4752,
|
52 |
+
"eval_gen_len": 14.6014,
|
53 |
+
"eval_jit_compilation_time": 12.9294,
|
54 |
+
"eval_loss": 2.9099602699279785,
|
55 |
+
"eval_runtime": 12.6411,
|
56 |
+
"eval_samples_per_second": 11.312,
|
57 |
+
"eval_steps_per_second": 2.848,
|
58 |
+
"step": 572
|
59 |
+
}
|
60 |
+
],
|
61 |
+
"max_steps": 572,
|
62 |
+
"num_train_epochs": 4,
|
63 |
+
"total_flos": 11456058359808.0,
|
64 |
+
"trial_name": null,
|
65 |
+
"trial_params": null
|
66 |
+
}
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:093ba43f3bf932ee2c0a0909229582d9d4b5b60aa73b0d1fe5caf0a48ad77ddb
|
3 |
+
size 4027
|