EhsanAghazadeh commited on
Commit
48a6ccf
1 Parent(s): b83cb11

Wrong model fixed.

Browse files
Files changed (6) hide show
  1. config.json +0 -1
  2. optimizer.pt +1 -1
  3. pytorch_model.bin +1 -1
  4. scheduler.pt +1 -1
  5. trainer_state.json +43 -13
  6. training_args.bin +1 -1
config.json CHANGED
@@ -14,7 +14,6 @@
14
  "end_n_top": 5,
15
  "eos_token_id": 2,
16
  "ff_activation": "gelu",
17
- "finetuning_task": "cola",
18
  "initializer_range": 0.02,
19
  "layer_norm_eps": 1e-12,
20
  "mem_len": null,
 
14
  "end_n_top": 5,
15
  "eos_token_id": 2,
16
  "ff_activation": "gelu",
 
17
  "initializer_range": 0.02,
18
  "layer_norm_eps": 1e-12,
19
  "mem_len": null,
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:afd8764f9775f328ceb63db680a192519cecbf130ccb101578c66fd7e6a93214
3
  size 2890816569
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fae204010e3ff30cf921d183114b85f1025c2ab9ff324d902627db2e0fa87d10
3
  size 2890816569
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:95b90a0014f1c8bd8eb44d3769b952d1af72e1906fe61dfcbd03a9577a2458b9
3
  size 1445440509
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a9ccb551009cac9035706f548db80c52424eee69e8cd58499dbf111a775aa7b
3
  size 1445440509
scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b10bbf9dd65fadb3be841f857dae7da43c92d17cbfe1737b283579e0bfca56e1
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1acca92a8b10bd06cc866eafb8aeb9d4f85bcdb2d57143bebcfe2bc573e5d33a
3
  size 623
trainer_state.json CHANGED
@@ -1,28 +1,58 @@
1
  {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 0.9354536950420954,
5
- "global_step": 1000,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
- "epoch": 0.47,
12
- "learning_rate": 4.4153414405986905e-05,
13
- "loss": 0.6437,
14
  "step": 500
15
  },
16
  {
17
- "epoch": 0.94,
18
- "learning_rate": 3.8306828811973814e-05,
19
- "loss": 0.6233,
 
 
 
 
 
 
 
 
20
  "step": 1000
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
21
  }
22
  ],
23
- "max_steps": 4276,
24
- "num_train_epochs": 4,
25
- "total_flos": 2219952844800000.0,
26
  "trial_name": null,
27
  "trial_params": null
28
  }
 
1
  {
2
+ "best_metric": 0.5940754350972381,
3
+ "best_model_checkpoint": "test-glue/checkpoint-1605",
4
+ "epoch": 3.0,
5
+ "global_step": 1605,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
+ "epoch": 0.93,
12
+ "learning_rate": 1.3769470404984425e-05,
13
+ "loss": 0.5592,
14
  "step": 500
15
  },
16
  {
17
+ "epoch": 1.0,
18
+ "eval_loss": 0.6476718187332153,
19
+ "eval_matthews_correlation": 0.30642694676082244,
20
+ "eval_runtime": 5.7468,
21
+ "eval_samples_per_second": 181.491,
22
+ "step": 535
23
+ },
24
+ {
25
+ "epoch": 1.87,
26
+ "learning_rate": 7.538940809968847e-06,
27
+ "loss": 0.4061,
28
  "step": 1000
29
+ },
30
+ {
31
+ "epoch": 2.0,
32
+ "eval_loss": 0.5062028169631958,
33
+ "eval_matthews_correlation": 0.5625644321063499,
34
+ "eval_runtime": 5.7304,
35
+ "eval_samples_per_second": 182.011,
36
+ "step": 1070
37
+ },
38
+ {
39
+ "epoch": 2.8,
40
+ "learning_rate": 1.308411214953271e-06,
41
+ "loss": 0.2545,
42
+ "step": 1500
43
+ },
44
+ {
45
+ "epoch": 3.0,
46
+ "eval_loss": 0.6555218696594238,
47
+ "eval_matthews_correlation": 0.5940754350972381,
48
+ "eval_runtime": 5.7265,
49
+ "eval_samples_per_second": 182.137,
50
+ "step": 1605
51
  }
52
  ],
53
+ "max_steps": 1605,
54
+ "num_train_epochs": 3,
55
+ "total_flos": 1169973683122500.0,
56
  "trial_name": null,
57
  "trial_params": null
58
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:75e62e34df944aa90b2bbc0df5665e61e488891c160d044d82d294a3c70b9a07
3
  size 2351
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f25199f88ba4cf62cc75f38680846823eeca9cfbd8160fb06568e3a992a63812
3
  size 2351