AlekseyKorshuk commited on
Commit
f17b1c6
1 Parent(s): 54b0783

huggingartists

Browse files
README.md CHANGED
@@ -45,15 +45,15 @@ from datasets import load_dataset
45
  dataset = load_dataset("huggingartists/zemfira")
46
  ```
47
 
48
- [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/fzlwlb18/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 Земфира (Zemfira)'s lyrics.
53
 
54
- Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1uuirro7) for full transparency and reproducibility.
55
 
56
- At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1uuirro7/artifacts) is logged and versioned.
57
 
58
  ## How to use
59
 
 
45
  dataset = load_dataset("huggingartists/zemfira")
46
  ```
47
 
48
+ [Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/3hj4sma8/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 Земфира (Zemfira)'s lyrics.
53
 
54
+ Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/1v74giz2) for full transparency and reproducibility.
55
 
56
+ At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/1v74giz2/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"
 
1
  {
2
+ "_name_or_path": "huggingartists/zemfira",
3
  "activation_function": "gelu_new",
4
  "architectures": [
5
  "GPT2LMHeadModel"
evaluation.txt CHANGED
@@ -1 +1 @@
1
- {"eval_loss": 1.7723504304885864, "eval_runtime": 3.5108, "eval_samples_per_second": 22.217, "eval_steps_per_second": 2.848, "epoch": 1.0}
 
1
+ {"eval_loss": 1.7192894220352173, "eval_runtime": 3.4091, "eval_samples_per_second": 20.827, "eval_steps_per_second": 2.64, "epoch": 2.0}
flax_model.msgpack CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e6459e8282ae04d405f178518877a3524a39a57cf65d9b2b5234a04a2694358c
3
  size 497764120
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c1f24d5f5d97214ea17828354e30c1cb0ab42893879b8f210bb789754e47b65
3
  size 497764120
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:dbd6c9f3b863771276c9e277199569516911bf865d02acb367349eb4619a9423
3
  size 995603825
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb2b4d0752c13f5268ee4193ab28152e257cd319d7febad402156ab94ce40514
3
  size 995603825
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:15862618e458991c730eb040418ef8bb80dc4a6e9c912d5031be31d5f793d819
3
  size 510403817
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2c544abb49c657e56c6e5d83148da11b08271b093394d03e86ca288f0a414cf
3
  size 510403817
rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:83944b113493814ae5c9c809188afdcbb000becae281a11123c189b2d376c1e0
3
- size 14503
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed9526fd15362f8845d61d83def08176b3bae047f567421fef1efe555a551faa
3
+ size 14567
scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7b0c69c25efb33387eda1137552cf03b56244a07541d58b5090744f9d9078bbc
3
  size 623
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a232bfd4c68f2b7c5d46c17b697d266203cb8334a94261d8334dd1f3e92002d
3
  size 623
tokenizer_config.json CHANGED
@@ -1 +1 @@
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
+ {"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": "huggingartists/zemfira", "tokenizer_class": "GPT2Tokenizer"}
trainer_state.json CHANGED
@@ -1,78 +1,152 @@
1
  {
2
- "best_metric": 1.7723504304885864,
3
- "best_model_checkpoint": "output/zemfira/checkpoint-49",
4
- "epoch": 1.0,
5
- "global_step": 49,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.1,
12
- "learning_rate": 0.00013370522424493186,
13
- "loss": 2.4725,
14
  "step": 5
15
  },
16
  {
17
  "epoch": 0.2,
18
- "learning_rate": 0.00012357697446014183,
19
- "loss": 2.1691,
20
  "step": 10
21
  },
22
  {
23
- "epoch": 0.31,
24
- "learning_rate": 0.00010784720288438084,
25
- "loss": 1.9883,
26
  "step": 15
27
  },
28
  {
29
- "epoch": 0.41,
30
- "learning_rate": 8.811859244288883e-05,
31
- "loss": 1.8517,
32
  "step": 20
33
  },
34
  {
35
- "epoch": 0.51,
36
- "learning_rate": 6.640126177858445e-05,
37
- "loss": 1.9739,
38
  "step": 25
39
  },
40
  {
41
- "epoch": 0.61,
42
- "learning_rate": 4.4907957266698305e-05,
43
- "loss": 1.9488,
44
  "step": 30
45
  },
46
  {
47
- "epoch": 0.71,
48
- "learning_rate": 2.5828599592490882e-05,
49
- "loss": 1.9148,
50
  "step": 35
51
  },
52
  {
53
- "epoch": 0.82,
54
- "learning_rate": 1.1107156004419736e-05,
55
- "loss": 1.9426,
56
  "step": 40
57
  },
58
  {
59
- "epoch": 0.92,
60
- "learning_rate": 2.24357239552258e-06,
61
- "loss": 2.0432,
62
  "step": 45
63
  },
64
  {
65
  "epoch": 1.0,
66
- "eval_loss": 1.7723504304885864,
67
- "eval_runtime": 3.4368,
68
- "eval_samples_per_second": 22.696,
69
- "eval_steps_per_second": 2.91,
70
- "step": 49
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
71
  }
72
  ],
73
- "max_steps": 49,
74
- "num_train_epochs": 1,
75
- "total_flos": 50821300224000.0,
76
  "trial_name": null,
77
  "trial_params": null
78
  }
 
1
  {
2
+ "best_metric": 1.7192894220352173,
3
+ "best_model_checkpoint": "output/zemfira/checkpoint-100",
4
+ "epoch": 2.0,
5
+ "global_step": 100,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.1,
12
+ "learning_rate": 3.357522982152468e-06,
13
+ "loss": 1.9295,
14
  "step": 5
15
  },
16
  {
17
  "epoch": 0.2,
18
+ "learning_rate": 1.3101434185878598e-05,
19
+ "loss": 1.7553,
20
  "step": 10
21
  },
22
  {
23
+ "epoch": 0.3,
24
+ "learning_rate": 2.8277931692736335e-05,
25
+ "loss": 1.9311,
26
  "step": 15
27
  },
28
  {
29
+ "epoch": 0.4,
30
+ "learning_rate": 4.74014341858786e-05,
31
+ "loss": 1.854,
32
  "step": 20
33
  },
34
  {
35
+ "epoch": 0.5,
36
+ "learning_rate": 6.859999999999999e-05,
37
+ "loss": 1.8097,
38
  "step": 25
39
  },
40
  {
41
+ "epoch": 0.6,
42
+ "learning_rate": 8.979856581412138e-05,
43
+ "loss": 1.9139,
44
  "step": 30
45
  },
46
  {
47
+ "epoch": 0.7,
48
+ "learning_rate": 0.00010892206830726364,
49
+ "loss": 1.8094,
50
  "step": 35
51
  },
52
  {
53
+ "epoch": 0.8,
54
+ "learning_rate": 0.0001240985658141214,
55
+ "loss": 1.6989,
56
  "step": 40
57
  },
58
  {
59
+ "epoch": 0.9,
60
+ "learning_rate": 0.00013384247701784754,
61
+ "loss": 1.7624,
62
  "step": 45
63
  },
64
  {
65
  "epoch": 1.0,
66
+ "learning_rate": 0.0001372,
67
+ "loss": 1.8189,
68
+ "step": 50
69
+ },
70
+ {
71
+ "epoch": 1.0,
72
+ "eval_loss": 1.7890528440475464,
73
+ "eval_runtime": 3.3356,
74
+ "eval_samples_per_second": 21.286,
75
+ "eval_steps_per_second": 2.698,
76
+ "step": 50
77
+ },
78
+ {
79
+ "epoch": 1.1,
80
+ "learning_rate": 0.00013384247701784754,
81
+ "loss": 1.7241,
82
+ "step": 55
83
+ },
84
+ {
85
+ "epoch": 1.2,
86
+ "learning_rate": 0.00012409856581412136,
87
+ "loss": 1.7957,
88
+ "step": 60
89
+ },
90
+ {
91
+ "epoch": 1.3,
92
+ "learning_rate": 0.00010892206830726372,
93
+ "loss": 1.7292,
94
+ "step": 65
95
+ },
96
+ {
97
+ "epoch": 1.4,
98
+ "learning_rate": 8.979856581412141e-05,
99
+ "loss": 1.6565,
100
+ "step": 70
101
+ },
102
+ {
103
+ "epoch": 1.5,
104
+ "learning_rate": 6.860000000000001e-05,
105
+ "loss": 1.6323,
106
+ "step": 75
107
+ },
108
+ {
109
+ "epoch": 1.6,
110
+ "learning_rate": 4.7401434185878625e-05,
111
+ "loss": 1.6913,
112
+ "step": 80
113
+ },
114
+ {
115
+ "epoch": 1.7,
116
+ "learning_rate": 2.827793169273636e-05,
117
+ "loss": 1.6236,
118
+ "step": 85
119
+ },
120
+ {
121
+ "epoch": 1.8,
122
+ "learning_rate": 1.310143418587862e-05,
123
+ "loss": 1.6434,
124
+ "step": 90
125
+ },
126
+ {
127
+ "epoch": 1.9,
128
+ "learning_rate": 3.3575229821524754e-06,
129
+ "loss": 1.7407,
130
+ "step": 95
131
+ },
132
+ {
133
+ "epoch": 2.0,
134
+ "learning_rate": 0.0,
135
+ "loss": 1.7433,
136
+ "step": 100
137
+ },
138
+ {
139
+ "epoch": 2.0,
140
+ "eval_loss": 1.7192894220352173,
141
+ "eval_runtime": 3.3307,
142
+ "eval_samples_per_second": 21.317,
143
+ "eval_steps_per_second": 2.702,
144
+ "step": 100
145
  }
146
  ],
147
+ "max_steps": 100,
148
+ "num_train_epochs": 2,
149
+ "total_flos": 103471644672000.0,
150
  "trial_name": null,
151
  "trial_params": null
152
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c540afccee3431a837d2b3268ea08f3ec4964cf38f2eac80dc5459e241364e15
3
  size 2671
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2e54519d4c930d1b7979099cf1dd3b0fcf472737a97a111bd2ce4bc6adb9b9e
3
  size 2671