AlekseyKorshuk
commited on
Commit
•
a0db3de
1
Parent(s):
6a0208e
huggingartists
Browse files- README.md +4 -4
- config.json +2 -2
- evaluation.txt +1 -1
- flax_model.msgpack +1 -1
- optimizer.pt +1 -1
- pytorch_model.bin +2 -2
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- tokenizer.json +0 -0
- tokenizer_config.json +1 -1
- trainer_state.json +113 -33
- training_args.bin +2 -2
README.md
CHANGED
@@ -14,7 +14,7 @@ widget:
|
|
14 |
<div class="inline-flex flex-col" style="line-height: 1.5;">
|
15 |
<div class="flex">
|
16 |
<div
|
17 |
-
style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://images.genius.com/
|
18 |
</div>
|
19 |
</div>
|
20 |
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
|
@@ -45,15 +45,15 @@ from datasets import load_dataset
|
|
45 |
dataset = load_dataset("huggingartists/rammstein")
|
46 |
```
|
47 |
|
48 |
-
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/
|
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 Rammstein's lyrics.
|
53 |
|
54 |
-
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/
|
55 |
|
56 |
-
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/
|
57 |
|
58 |
## How to use
|
59 |
|
14 |
<div class="inline-flex flex-col" style="line-height: 1.5;">
|
15 |
<div class="flex">
|
16 |
<div
|
17 |
+
style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://images.genius.com/29cedf8dd30a7458f4fca47d1c0f0eab.1000x1000x1.jpg')">
|
18 |
</div>
|
19 |
</div>
|
20 |
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
|
45 |
dataset = load_dataset("huggingartists/rammstein")
|
46 |
```
|
47 |
|
48 |
+
[Explore the data](https://wandb.ai/huggingartists/huggingartists/runs/qt3qa1x1/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 Rammstein's lyrics.
|
53 |
|
54 |
+
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/huggingartists/huggingartists/runs/2yyigjzv) for full transparency and reproducibility.
|
55 |
|
56 |
+
At the end of training, [the final model](https://wandb.ai/huggingartists/huggingartists/runs/2yyigjzv/artifacts) is logged and versioned.
|
57 |
|
58 |
## How to use
|
59 |
|
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_function": "gelu_new",
|
4 |
"architectures": [
|
5 |
"GPT2LMHeadModel"
|
@@ -36,7 +36,7 @@
|
|
36 |
}
|
37 |
},
|
38 |
"torch_dtype": "float32",
|
39 |
-
"transformers_version": "4.
|
40 |
"use_cache": true,
|
41 |
"vocab_size": 50257
|
42 |
}
|
1 |
{
|
2 |
+
"_name_or_path": "rammstein",
|
3 |
"activation_function": "gelu_new",
|
4 |
"architectures": [
|
5 |
"GPT2LMHeadModel"
|
36 |
}
|
37 |
},
|
38 |
"torch_dtype": "float32",
|
39 |
+
"transformers_version": "4.19.2",
|
40 |
"use_cache": true,
|
41 |
"vocab_size": 50257
|
42 |
}
|
evaluation.txt
CHANGED
@@ -1 +1 @@
|
|
1 |
-
{"eval_loss": 2.
|
1 |
+
{"eval_loss": 2.3011960983276367, "eval_runtime": 0.7541, "eval_samples_per_second": 79.564, "eval_steps_per_second": 10.609, "epoch": 2.0}
|
flax_model.msgpack
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 497764120
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:03c08e945202ddf1ebd8348598751d6c196aa56293a3859e6127461b5c6be2ad
|
3 |
size 497764120
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 995603825
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cebc41cae62202846215c2c1720dc6e09f5c67ecb6bfadaf587db41eb94d3086
|
3 |
size 995603825
|
pytorch_model.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:e338ab740793bb9a9e4aac9c5c91021b616ca47dba2e7f98c96b90f92da75a93
|
3 |
+
size 510396521
|
rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14503
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5fa5222bb5aaf4570a2485f42de06ccf0e38362fb68506b4fa8d9280490a1a50
|
3 |
size 14503
|
scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:12ce1132cbb6aa5c18263e33796f9990b4148f45c42e80bbd2185f1d18ecd3d7
|
3 |
size 623
|
tokenizer.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
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": "
|
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/rammstein", "tokenizer_class": "GPT2Tokenizer"}
|
trainer_state.json
CHANGED
@@ -1,66 +1,146 @@
|
|
1 |
{
|
2 |
-
"best_metric": 2.
|
3 |
-
"best_model_checkpoint": "output/rammstein/checkpoint-
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
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.
|
12 |
-
"learning_rate":
|
13 |
-
"loss":
|
14 |
"step": 5
|
15 |
},
|
16 |
{
|
17 |
-
"epoch": 0.
|
18 |
-
"learning_rate":
|
19 |
-
"loss":
|
20 |
"step": 10
|
21 |
},
|
22 |
{
|
23 |
-
"epoch": 0.
|
24 |
-
"learning_rate":
|
25 |
-
"loss": 3.
|
26 |
"step": 15
|
27 |
},
|
28 |
{
|
29 |
-
"epoch": 0.
|
30 |
-
"learning_rate": 5.
|
31 |
-
"loss": 2.
|
32 |
"step": 20
|
33 |
},
|
34 |
{
|
35 |
-
"epoch": 0.
|
36 |
-
"learning_rate":
|
37 |
-
"loss": 2.
|
38 |
"step": 25
|
39 |
},
|
40 |
{
|
41 |
-
"epoch": 0.
|
42 |
-
"learning_rate": 9.
|
43 |
-
"loss":
|
44 |
"step": 30
|
45 |
},
|
46 |
{
|
47 |
-
"epoch": 0.
|
48 |
-
"learning_rate":
|
49 |
-
"loss":
|
50 |
"step": 35
|
51 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
{
|
53 |
"epoch": 1.0,
|
54 |
-
"eval_loss": 2.
|
55 |
-
"eval_runtime":
|
56 |
-
"eval_samples_per_second":
|
57 |
-
"eval_steps_per_second":
|
58 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
}
|
60 |
],
|
61 |
-
"max_steps":
|
62 |
-
"num_train_epochs":
|
63 |
-
"total_flos":
|
64 |
"trial_name": null,
|
65 |
"trial_params": null
|
66 |
}
|
1 |
{
|
2 |
+
"best_metric": 2.3011960983276367,
|
3 |
+
"best_model_checkpoint": "output/rammstein/checkpoint-96",
|
4 |
+
"epoch": 2.0,
|
5 |
+
"global_step": 96,
|
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": 3.64059311663575e-06,
|
13 |
+
"loss": 2.9482,
|
14 |
"step": 5
|
15 |
},
|
16 |
{
|
17 |
+
"epoch": 0.21,
|
18 |
+
"learning_rate": 1.4175960856021253e-05,
|
19 |
+
"loss": 2.7279,
|
20 |
"step": 10
|
21 |
},
|
22 |
{
|
23 |
+
"epoch": 0.31,
|
24 |
+
"learning_rate": 3.048788201485529e-05,
|
25 |
+
"loss": 3.0408,
|
26 |
"step": 15
|
27 |
},
|
28 |
{
|
29 |
+
"epoch": 0.42,
|
30 |
+
"learning_rate": 5.084501350596709e-05,
|
31 |
+
"loss": 2.8174,
|
32 |
"step": 20
|
33 |
},
|
34 |
{
|
35 |
+
"epoch": 0.52,
|
36 |
+
"learning_rate": 7.308665466518777e-05,
|
37 |
+
"loss": 2.9266,
|
38 |
"step": 25
|
39 |
},
|
40 |
{
|
41 |
+
"epoch": 0.62,
|
42 |
+
"learning_rate": 9.485208346024518e-05,
|
43 |
+
"loss": 2.7754,
|
44 |
"step": 30
|
45 |
},
|
46 |
{
|
47 |
+
"epoch": 0.73,
|
48 |
+
"learning_rate": 0.00011383112291586474,
|
49 |
+
"loss": 2.4825,
|
50 |
"step": 35
|
51 |
},
|
52 |
+
{
|
53 |
+
"epoch": 0.83,
|
54 |
+
"learning_rate": 0.00012800934269961248,
|
55 |
+
"loss": 2.6286,
|
56 |
+
"step": 40
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.94,
|
60 |
+
"learning_rate": 0.0001358818702356616,
|
61 |
+
"loss": 2.681,
|
62 |
+
"step": 45
|
63 |
+
},
|
64 |
{
|
65 |
"epoch": 1.0,
|
66 |
+
"eval_loss": 2.4657435417175293,
|
67 |
+
"eval_runtime": 0.7344,
|
68 |
+
"eval_samples_per_second": 81.697,
|
69 |
+
"eval_steps_per_second": 10.893,
|
70 |
+
"step": 48
|
71 |
+
},
|
72 |
+
{
|
73 |
+
"epoch": 1.04,
|
74 |
+
"learning_rate": 0.0001366131174902434,
|
75 |
+
"loss": 2.5211,
|
76 |
+
"step": 50
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 1.15,
|
80 |
+
"learning_rate": 0.0001301254700691424,
|
81 |
+
"loss": 2.3708,
|
82 |
+
"step": 55
|
83 |
+
},
|
84 |
+
{
|
85 |
+
"epoch": 1.25,
|
86 |
+
"learning_rate": 0.00011710752518939717,
|
87 |
+
"loss": 2.4006,
|
88 |
+
"step": 60
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.35,
|
92 |
+
"learning_rate": 9.894100414902355e-05,
|
93 |
+
"loss": 2.3993,
|
94 |
+
"step": 65
|
95 |
+
},
|
96 |
+
{
|
97 |
+
"epoch": 1.46,
|
98 |
+
"learning_rate": 7.755409678629555e-05,
|
99 |
+
"loss": 2.4798,
|
100 |
+
"step": 70
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 1.56,
|
104 |
+
"learning_rate": 5.5216803909693664e-05,
|
105 |
+
"loss": 2.4973,
|
106 |
+
"step": 75
|
107 |
+
},
|
108 |
+
{
|
109 |
+
"epoch": 1.67,
|
110 |
+
"learning_rate": 3.4300000000000054e-05,
|
111 |
+
"loss": 2.2237,
|
112 |
+
"step": 80
|
113 |
+
},
|
114 |
+
{
|
115 |
+
"epoch": 1.77,
|
116 |
+
"learning_rate": 1.7023789206942107e-05,
|
117 |
+
"loss": 2.2823,
|
118 |
+
"step": 85
|
119 |
+
},
|
120 |
+
{
|
121 |
+
"epoch": 1.88,
|
122 |
+
"learning_rate": 5.221864069725753e-06,
|
123 |
+
"loss": 2.2183,
|
124 |
+
"step": 90
|
125 |
+
},
|
126 |
+
{
|
127 |
+
"epoch": 1.98,
|
128 |
+
"learning_rate": 1.4687786583180135e-07,
|
129 |
+
"loss": 2.2238,
|
130 |
+
"step": 95
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 2.0,
|
134 |
+
"eval_loss": 2.3011960983276367,
|
135 |
+
"eval_runtime": 0.7368,
|
136 |
+
"eval_samples_per_second": 81.434,
|
137 |
+
"eval_steps_per_second": 10.858,
|
138 |
+
"step": 96
|
139 |
}
|
140 |
],
|
141 |
+
"max_steps": 96,
|
142 |
+
"num_train_epochs": 2,
|
143 |
+
"total_flos": 99552264192000.0,
|
144 |
"trial_name": null,
|
145 |
"trial_params": null
|
146 |
}
|
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 sha256:21bd930efff238a97f2bba32a2ac5d18401b5fa862717f400510445f4a9addf4
|
3 |
+
size 3247
|