AlekseyKorshuk commited on
Commit
a806d4d
1 Parent(s): bc37edc

huggingartists

Browse files
README.md CHANGED
@@ -45,15 +45,15 @@ from datasets import load_dataset
45
  dataset = load_dataset("huggingartists/gspd")
46
  ```
47
 
48
- [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/17su8b0l/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
49
 
50
  ## Training procedure
51
 
52
  The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on GSPD's lyrics.
53
 
54
- Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/cnw6rr26) for full transparency and reproducibility.
55
 
56
- At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/cnw6rr26/artifacts) is logged and versioned.
57
 
58
  ## How to use
59
 
45
  dataset = load_dataset("huggingartists/gspd")
46
  ```
47
 
48
+ [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/3jof0sex/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
49
 
50
  ## Training procedure
51
 
52
  The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on GSPD's lyrics.
53
 
54
+ Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/2nxhrny4) for full transparency and reproducibility.
55
 
56
+ At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/2nxhrny4/artifacts) is logged and versioned.
57
 
58
  ## How to use
59
 
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "gpt2",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
@@ -18,7 +18,9 @@
18
  "n_inner": null,
19
  "n_layer": 12,
20
  "n_positions": 1024,
 
21
  "resid_pdrop": 0.1,
 
22
  "scale_attn_weights": true,
23
  "summary_activation": null,
24
  "summary_first_dropout": 0.1,
@@ -35,7 +37,7 @@
35
  }
36
  },
37
  "torch_dtype": "float32",
38
- "transformers_version": "4.10.0",
39
  "use_cache": true,
40
  "vocab_size": 50257
41
  }
1
  {
2
+ "_name_or_path": "gspd",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
18
  "n_inner": null,
19
  "n_layer": 12,
20
  "n_positions": 1024,
21
+ "reorder_and_upcast_attn": false,
22
  "resid_pdrop": 0.1,
23
+ "scale_attn_by_inverse_layer_idx": false,
24
  "scale_attn_weights": true,
25
  "summary_activation": null,
26
  "summary_first_dropout": 0.1,
37
  }
38
  },
39
  "torch_dtype": "float32",
40
+ "transformers_version": "4.21.1",
41
  "use_cache": true,
42
  "vocab_size": 50257
43
  }
evaluation.txt CHANGED
@@ -1 +1 @@
1
- {"eval_loss": 1.6143786907196045, "eval_runtime": 6.648, "eval_samples_per_second": 22.563, "eval_steps_per_second": 2.858, "epoch": 1.0}
1
+ {"eval_loss": 1.6192371845245361, "eval_runtime": 3.877, "eval_samples_per_second": 44.106, "eval_steps_per_second": 5.674, "epoch": 2.0}
flax_model.msgpack CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:903a60b019c3189ea2401d825a51588abe0295c21127eadc7f170f86b4711695
3
  size 497764120
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d7317d7b2d5b31cf6792c6f9b1aae151024c566ab5d2cdad309b0d8787e2d3e
3
  size 497764120
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5bf9330ebc61aeb2214f532e874e79d806f5773424f220a2a22e25116ae7d7ab
3
- size 995603825
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db51e790050e6606874c3cbde9a56238c405ff67191cdab7b7eecb47d32f638a
3
+ size 995604017
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:238219c7329b0b10459c22d9cd9fd9efb44970d69676335af067af5088ae373d
3
- size 510403817
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:088ce82b45d839e12925669e84a76044f7f5dffaca9ddb8a53fb30425464187c
3
+ size 510396521
rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1faa9800090f58192e975d9de4af52a7d1d95073c9339bf79e7455df2e035748
3
  size 14567
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:609e3e95ba1d71e778b2f2f711f521fee32d49f7719314a7a6f9c7574f7692ec
3
  size 14567
scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d1c2f47e2da491a315f90603f1d3a2627d77a6f03326673b420719eb45024ca9
3
  size 623
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5643af0d2d309ea9f8d4e15ffbc8ebd3faca59b9cbbd8bf85f08e202e280408b
3
  size 623
special_tokens_map.json CHANGED
@@ -1 +1,5 @@
1
- {"bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "unk_token": "<|endoftext|>"}
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
tokenizer_config.json CHANGED
@@ -1 +1,10 @@
1
- {"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "gpt2", "tokenizer_class": "GPT2Tokenizer"}
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "eos_token": "<|endoftext|>",
5
+ "model_max_length": 1024,
6
+ "name_or_path": "huggingartists/gspd",
7
+ "special_tokens_map_file": null,
8
+ "tokenizer_class": "GPT2Tokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
trainer_state.json CHANGED
@@ -1,144 +1,144 @@
1
  {
2
- "best_metric": 1.6143786907196045,
3
- "best_model_checkpoint": "output/gspd/checkpoint-104",
4
  "epoch": 1.0,
5
- "global_step": 104,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.05,
12
- "learning_rate": 0.00013641901727743684,
13
- "loss": 2.3995,
14
  "step": 5
15
  },
16
  {
17
  "epoch": 0.1,
18
- "learning_rate": 0.00013409385144248624,
19
- "loss": 2.0772,
20
  "step": 10
21
  },
22
  {
23
- "epoch": 0.14,
24
- "learning_rate": 0.00013027744460430822,
25
- "loss": 1.9236,
26
  "step": 15
27
  },
28
  {
29
- "epoch": 0.19,
30
- "learning_rate": 0.00012505669320030482,
31
- "loss": 1.9836,
32
  "step": 20
33
  },
34
  {
35
- "epoch": 0.24,
36
- "learning_rate": 0.00011855046943587908,
37
- "loss": 1.876,
38
  "step": 25
39
  },
40
  {
41
  "epoch": 0.29,
42
- "learning_rate": 0.00011090691466231807,
43
- "loss": 1.9239,
44
  "step": 30
45
  },
46
  {
47
  "epoch": 0.34,
48
- "learning_rate": 0.00010230006632035399,
49
- "loss": 1.8628,
50
  "step": 35
51
  },
52
  {
53
- "epoch": 0.38,
54
- "learning_rate": 9.292589525111794e-05,
55
- "loss": 1.6981,
56
  "step": 40
57
  },
58
  {
59
- "epoch": 0.43,
60
- "learning_rate": 8.299784360164853e-05,
61
- "loss": 1.7632,
62
  "step": 45
63
  },
64
  {
65
- "epoch": 0.48,
66
- "learning_rate": 7.274196492316882e-05,
67
- "loss": 1.8217,
68
  "step": 50
69
  },
70
  {
71
- "epoch": 0.53,
72
- "learning_rate": 6.239177711808685e-05,
73
- "loss": 1.7681,
74
  "step": 55
75
  },
76
  {
77
- "epoch": 0.58,
78
- "learning_rate": 5.218294542987356e-05,
79
- "loss": 1.7394,
80
  "step": 60
81
  },
82
  {
83
- "epoch": 0.62,
84
- "learning_rate": 4.2347916539754844e-05,
85
- "loss": 1.8022,
86
  "step": 65
87
  },
88
  {
89
- "epoch": 0.67,
90
- "learning_rate": 3.311062594741274e-05,
91
- "loss": 1.7416,
92
  "step": 70
93
  },
94
  {
95
- "epoch": 0.72,
96
- "learning_rate": 2.4681399144273327e-05,
97
- "loss": 1.7103,
98
  "step": 75
99
  },
100
  {
101
- "epoch": 0.77,
102
- "learning_rate": 1.725216267546246e-05,
103
- "loss": 1.7463,
104
  "step": 80
105
  },
106
  {
107
- "epoch": 0.82,
108
- "learning_rate": 1.0992074130600507e-05,
109
- "loss": 1.6627,
110
  "step": 85
111
  },
112
  {
113
- "epoch": 0.87,
114
- "learning_rate": 6.043670564942622e-06,
115
- "loss": 1.7961,
116
  "step": 90
117
  },
118
  {
119
- "epoch": 0.91,
120
- "learning_rate": 2.519623048140383e-06,
121
- "loss": 1.7855,
122
  "step": 95
123
  },
124
  {
125
- "epoch": 0.96,
126
- "learning_rate": 5.001712368734975e-07,
127
- "loss": 1.7266,
128
  "step": 100
129
  },
130
  {
131
  "epoch": 1.0,
132
- "eval_loss": 1.6143786907196045,
133
- "eval_runtime": 6.5105,
134
- "eval_samples_per_second": 23.04,
135
- "eval_steps_per_second": 2.918,
136
- "step": 104
137
  }
138
  ],
139
- "max_steps": 104,
140
- "num_train_epochs": 1,
141
- "total_flos": 108566839296000.0,
142
  "trial_name": null,
143
  "trial_params": null
144
  }
1
  {
2
+ "best_metric": 1.6192371845245361,
3
+ "best_model_checkpoint": "output/gspd/checkpoint-102",
4
  "epoch": 1.0,
5
+ "global_step": 102,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.05,
12
+ "learning_rate": 0.00013638815138477438,
13
+ "loss": 1.4771,
14
  "step": 5
15
  },
16
  {
17
  "epoch": 0.1,
18
+ "learning_rate": 0.00013397182122930294,
19
+ "loss": 1.6873,
20
  "step": 10
21
  },
22
  {
23
+ "epoch": 0.15,
24
+ "learning_rate": 0.0001300082017869573,
25
+ "loss": 1.5944,
26
  "step": 15
27
  },
28
  {
29
+ "epoch": 0.2,
30
+ "learning_rate": 0.0001245911081876368,
31
+ "loss": 1.563,
32
  "step": 20
33
  },
34
  {
35
+ "epoch": 0.25,
36
+ "learning_rate": 0.0001178487579222207,
37
+ "loss": 1.5602,
38
  "step": 25
39
  },
40
  {
41
  "epoch": 0.29,
42
+ "learning_rate": 0.00010994073605561706,
43
+ "loss": 1.5302,
44
  "step": 30
45
  },
46
  {
47
  "epoch": 0.34,
48
+ "learning_rate": 0.00010105421799895028,
49
+ "loss": 1.5424,
50
  "step": 35
51
  },
52
  {
53
+ "epoch": 0.39,
54
+ "learning_rate": 9.139953924430467e-05,
55
+ "loss": 1.6088,
56
  "step": 40
57
  },
58
  {
59
+ "epoch": 0.44,
60
+ "learning_rate": 8.120521692221671e-05,
61
+ "loss": 1.6842,
62
  "step": 45
63
  },
64
  {
65
+ "epoch": 0.49,
66
+ "learning_rate": 7.071254101695329e-05,
67
+ "loss": 1.5729,
68
  "step": 50
69
  },
70
  {
71
+ "epoch": 0.54,
72
+ "learning_rate": 6.0169863260400656e-05,
73
+ "loss": 1.4031,
74
  "step": 55
75
  },
76
  {
77
+ "epoch": 0.59,
78
+ "learning_rate": 4.9826718881055135e-05,
79
+ "loss": 1.4452,
80
  "step": 60
81
  },
82
  {
83
+ "epoch": 0.64,
84
+ "learning_rate": 3.992792034076664e-05,
85
+ "loss": 1.6059,
86
  "step": 65
87
  },
88
  {
89
+ "epoch": 0.69,
90
+ "learning_rate": 3.0707762854909365e-05,
91
+ "loss": 1.4848,
92
  "step": 70
93
  },
94
  {
95
+ "epoch": 0.74,
96
+ "learning_rate": 2.2384478845846205e-05,
97
+ "loss": 1.531,
98
  "step": 75
99
  },
100
  {
101
+ "epoch": 0.78,
102
+ "learning_rate": 1.515507258753902e-05,
103
+ "loss": 1.4834,
104
  "step": 80
105
  },
106
  {
107
+ "epoch": 0.83,
108
+ "learning_rate": 9.190657300387498e-06,
109
+ "loss": 1.4458,
110
  "step": 85
111
  },
112
  {
113
+ "epoch": 0.88,
114
+ "learning_rate": 4.6324050628612214e-06,
115
+ "loss": 1.5666,
116
  "step": 90
117
  },
118
  {
119
+ "epoch": 0.93,
120
+ "learning_rate": 1.588205401691401e-06,
121
+ "loss": 1.5073,
122
  "step": 95
123
  },
124
  {
125
+ "epoch": 0.98,
126
+ "learning_rate": 1.3011164863877445e-07,
127
+ "loss": 1.4195,
128
  "step": 100
129
  },
130
  {
131
  "epoch": 1.0,
132
+ "eval_loss": 1.6192371845245361,
133
+ "eval_runtime": 3.8612,
134
+ "eval_samples_per_second": 44.287,
135
+ "eval_steps_per_second": 5.698,
136
+ "step": 102
137
  }
138
  ],
139
+ "max_steps": 204,
140
+ "num_train_epochs": 2,
141
+ "total_flos": 105823272960000.0,
142
  "trial_name": null,
143
  "trial_params": null
144
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:87567da31bf0b1dfd45bb9ff251c20e21f066cdbbeca02722787462d643334e0
3
- size 2671
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91a9ae3b4198707837d8729d85e9c1c901dba6204ce15dbd4033f413af824434
3
+ size 3375