Yoshisaur commited on
Commit
a1f51ea
1 Parent(s): c8408b2

lr=0.00001, epoch=500

Browse files
all_results.json CHANGED
@@ -1,17 +1,23 @@
1
  {
2
- "epoch": 3.0,
3
- "eval_loss": 3.4407730102539062,
4
- "eval_mem_cpu_alloc_delta": 29202,
5
- "eval_mem_cpu_peaked_delta": 71004,
6
- "eval_runtime": 66.3528,
7
- "eval_samples": 6,
8
- "eval_samples_per_second": 0.09,
9
- "init_mem_cpu_alloc_delta": 22589,
10
- "init_mem_cpu_peaked_delta": 2808,
11
- "perplexity": 31.211075326966256,
12
- "train_mem_cpu_alloc_delta": 263088,
13
- "train_mem_cpu_peaked_delta": 110881,
14
- "train_runtime": 4435.5996,
15
- "train_samples": 6,
16
- "train_samples_per_second": 0.004
 
 
 
 
 
 
17
  }
1
  {
2
+ "epoch": 500.0,
3
+ "eval_loss": 0.035903312265872955,
4
+ "eval_mem_cpu_alloc_delta": 72395,
5
+ "eval_mem_cpu_peaked_delta": 26654,
6
+ "eval_mem_gpu_alloc_delta": 0,
7
+ "eval_mem_gpu_peaked_delta": 379483648,
8
+ "eval_runtime": 0.5933,
9
+ "eval_samples": 17,
10
+ "eval_samples_per_second": 28.652,
11
+ "init_mem_cpu_alloc_delta": 66338,
12
+ "init_mem_cpu_peaked_delta": 19115,
13
+ "init_mem_gpu_alloc_delta": 454524928,
14
+ "init_mem_gpu_peaked_delta": 0,
15
+ "perplexity": 1.0365556194314987,
16
+ "train_mem_cpu_alloc_delta": 411153,
17
+ "train_mem_cpu_peaked_delta": 658045,
18
+ "train_mem_gpu_alloc_delta": 1347066368,
19
+ "train_mem_gpu_peaked_delta": 3584124928,
20
+ "train_runtime": 1015.5322,
21
+ "train_samples": 17,
22
+ "train_samples_per_second": 8.37
23
  }
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "rinna/japanese-gpt2-medium",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
@@ -13,10 +13,10 @@
13
  "layer_norm_epsilon": 1e-05,
14
  "model_type": "gpt2",
15
  "n_ctx": 1024,
16
- "n_embd": 1024,
17
- "n_head": 16,
18
- "n_inner": 4096,
19
- "n_layer": 24,
20
  "n_positions": 1024,
21
  "resid_pdrop": 0.1,
22
  "scale_attn_weights": true,
1
  {
2
+ "_name_or_path": "rinna/japanese-gpt2-small",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
13
  "layer_norm_epsilon": 1e-05,
14
  "model_type": "gpt2",
15
  "n_ctx": 1024,
16
+ "n_embd": 768,
17
+ "n_head": 12,
18
+ "n_inner": 3072,
19
+ "n_layer": 12,
20
  "n_positions": 1024,
21
  "resid_pdrop": 0.1,
22
  "scale_attn_weights": true,
eval_results.json CHANGED
@@ -1,10 +1,12 @@
1
  {
2
- "epoch": 3.0,
3
- "eval_loss": 3.4407730102539062,
4
- "eval_mem_cpu_alloc_delta": 29202,
5
- "eval_mem_cpu_peaked_delta": 71004,
6
- "eval_runtime": 66.3528,
7
- "eval_samples": 6,
8
- "eval_samples_per_second": 0.09,
9
- "perplexity": 31.211075326966256
 
 
10
  }
1
  {
2
+ "epoch": 500.0,
3
+ "eval_loss": 0.035903312265872955,
4
+ "eval_mem_cpu_alloc_delta": 72395,
5
+ "eval_mem_cpu_peaked_delta": 26654,
6
+ "eval_mem_gpu_alloc_delta": 0,
7
+ "eval_mem_gpu_peaked_delta": 379483648,
8
+ "eval_runtime": 0.5933,
9
+ "eval_samples": 17,
10
+ "eval_samples_per_second": 28.652,
11
+ "perplexity": 1.0365556194314987
12
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0a9dff447319bdafbae8b779b16f05c2da58d3453815d88855a1fdafae4fbdd7
3
- size 1369795865
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89de45eb957f3abc11ea3cbea0c81ef0835090b4ecb3bb6f4d09db7e817f7eb3
3
+ size 454318313
tokenizer_config.json CHANGED
@@ -1 +1 @@
1
- {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "[PAD]", "extra_ids": 0, "additional_special_tokens": [], "sp_model_kwargs": {}, "bos_token": "<s>", "cls_token": "[CLS]", "sep_token": "[SEP]", "mask_token": "[MASK]", "do_lower_case": true, "tokenizer_class": "T5Tokenizer", "use_fast": false, "special_tokens_map_file": "/Users/yoshisaur/.cache/huggingface/transformers/42091916a8a40b3949b8a4f56ce63e437a166ae0e88d1d15546860c13bdc5ceb.9049458ebcd1cf666b7b0a046aa394597f12e611077571cfc86e0938f8675d82", "tokenizer_file": null, "name_or_path": "rinna/japanese-gpt2-medium"}
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "[PAD]", "extra_ids": 0, "additional_special_tokens": [], "sp_model_kwargs": {}, "bos_token": "<s>", "cls_token": "[CLS]", "sep_token": "[SEP]", "mask_token": "[MASK]", "do_lower_case": true, "tokenizer_class": "T5Tokenizer", "use_fast": false, "special_tokens_map_file": "/home/student/e20/e205723/.cache/huggingface/transformers/96f1ddf7247675414b491c5c3ae4c7ac307cf28d0f90977b7cbdf13a8f84bfc8.9049458ebcd1cf666b7b0a046aa394597f12e611077571cfc86e0938f8675d82", "tokenizer_file": null, "name_or_path": "rinna/japanese-gpt2-small"}
train_results.json CHANGED
@@ -1,10 +1,14 @@
1
  {
2
- "epoch": 3.0,
3
- "init_mem_cpu_alloc_delta": 22589,
4
- "init_mem_cpu_peaked_delta": 2808,
5
- "train_mem_cpu_alloc_delta": 263088,
6
- "train_mem_cpu_peaked_delta": 110881,
7
- "train_runtime": 4435.5996,
8
- "train_samples": 6,
9
- "train_samples_per_second": 0.004
 
 
 
 
10
  }
1
  {
2
+ "epoch": 500.0,
3
+ "init_mem_cpu_alloc_delta": 66338,
4
+ "init_mem_cpu_peaked_delta": 19115,
5
+ "init_mem_gpu_alloc_delta": 454524928,
6
+ "init_mem_gpu_peaked_delta": 0,
7
+ "train_mem_cpu_alloc_delta": 411153,
8
+ "train_mem_cpu_peaked_delta": 658045,
9
+ "train_mem_gpu_alloc_delta": 1347066368,
10
+ "train_mem_gpu_peaked_delta": 3584124928,
11
+ "train_runtime": 1015.5322,
12
+ "train_samples": 17,
13
+ "train_samples_per_second": 8.37
14
  }
trainer_state.json CHANGED
@@ -1,23 +1,125 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 3.0,
5
- "global_step": 18,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
- "epoch": 3.0,
12
- "step": 18,
13
- "total_flos": 37173067776000.0,
14
- "train_runtime": 4435.5996,
15
- "train_samples_per_second": 0.004
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
  }
17
  ],
18
- "max_steps": 18,
19
- "num_train_epochs": 3,
20
- "total_flos": 37173067776000.0,
21
  "trial_name": null,
22
  "trial_params": null
23
  }
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 500.0,
5
+ "global_step": 8500,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
+ "epoch": 29.41,
12
+ "learning_rate": 9.411764705882354e-06,
13
+ "loss": 3.2581,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 58.82,
18
+ "learning_rate": 8.823529411764707e-06,
19
+ "loss": 1.3987,
20
+ "step": 1000
21
+ },
22
+ {
23
+ "epoch": 88.24,
24
+ "learning_rate": 8.23529411764706e-06,
25
+ "loss": 0.6075,
26
+ "step": 1500
27
+ },
28
+ {
29
+ "epoch": 117.65,
30
+ "learning_rate": 7.647058823529411e-06,
31
+ "loss": 0.2936,
32
+ "step": 2000
33
+ },
34
+ {
35
+ "epoch": 147.06,
36
+ "learning_rate": 7.058823529411766e-06,
37
+ "loss": 0.1595,
38
+ "step": 2500
39
+ },
40
+ {
41
+ "epoch": 176.47,
42
+ "learning_rate": 6.470588235294119e-06,
43
+ "loss": 0.1009,
44
+ "step": 3000
45
+ },
46
+ {
47
+ "epoch": 205.88,
48
+ "learning_rate": 5.882352941176471e-06,
49
+ "loss": 0.0765,
50
+ "step": 3500
51
+ },
52
+ {
53
+ "epoch": 235.29,
54
+ "learning_rate": 5.294117647058824e-06,
55
+ "loss": 0.0646,
56
+ "step": 4000
57
+ },
58
+ {
59
+ "epoch": 264.71,
60
+ "learning_rate": 4.705882352941177e-06,
61
+ "loss": 0.058,
62
+ "step": 4500
63
+ },
64
+ {
65
+ "epoch": 294.12,
66
+ "learning_rate": 4.11764705882353e-06,
67
+ "loss": 0.0533,
68
+ "step": 5000
69
+ },
70
+ {
71
+ "epoch": 323.53,
72
+ "learning_rate": 3.529411764705883e-06,
73
+ "loss": 0.0501,
74
+ "step": 5500
75
+ },
76
+ {
77
+ "epoch": 352.94,
78
+ "learning_rate": 2.9411764705882355e-06,
79
+ "loss": 0.0477,
80
+ "step": 6000
81
+ },
82
+ {
83
+ "epoch": 382.35,
84
+ "learning_rate": 2.3529411764705885e-06,
85
+ "loss": 0.046,
86
+ "step": 6500
87
+ },
88
+ {
89
+ "epoch": 411.76,
90
+ "learning_rate": 1.7647058823529414e-06,
91
+ "loss": 0.0449,
92
+ "step": 7000
93
+ },
94
+ {
95
+ "epoch": 441.18,
96
+ "learning_rate": 1.1764705882352942e-06,
97
+ "loss": 0.0439,
98
+ "step": 7500
99
+ },
100
+ {
101
+ "epoch": 470.59,
102
+ "learning_rate": 5.882352941176471e-07,
103
+ "loss": 0.0435,
104
+ "step": 8000
105
+ },
106
+ {
107
+ "epoch": 500.0,
108
+ "learning_rate": 0.0,
109
+ "loss": 0.0429,
110
+ "step": 8500
111
+ },
112
+ {
113
+ "epoch": 500.0,
114
+ "step": 8500,
115
+ "total_flos": 5766492192768000.0,
116
+ "train_runtime": 1015.5322,
117
+ "train_samples_per_second": 8.37
118
  }
119
  ],
120
+ "max_steps": 8500,
121
+ "num_train_epochs": 500,
122
+ "total_flos": 5766492192768000.0,
123
  "trial_name": null,
124
  "trial_params": null
125
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4f62138ffda72cf15563c12224eac8f955ca016c11ee533dd99069a4b9a73cb9
3
- size 2287
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19f51ac4b961f84e7552e792084e2b41ede999e808a62010fb8eac8491fb617d
3
+ size 2223