jialicheng
commited on
Commit
•
bd39069
1
Parent(s):
cb8af26
Upload folder using huggingface_hub
Browse files- README.md +86 -0
- all_results.json +19 -0
- config.json +181 -0
- eval_results.json +8 -0
- model.safetensors +3 -0
- pred_logit_eval.npy +3 -0
- pred_logit_test.npy +3 -0
- pred_logit_train.npy +3 -0
- preprocessor_config.json +14 -0
- test_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +225 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
base_model: openai/whisper-small
|
4 |
+
tags:
|
5 |
+
- audio-classification
|
6 |
+
- generated_from_trainer
|
7 |
+
datasets:
|
8 |
+
- superb
|
9 |
+
metrics:
|
10 |
+
- accuracy
|
11 |
+
model-index:
|
12 |
+
- name: superb_ks_42
|
13 |
+
results:
|
14 |
+
- task:
|
15 |
+
name: Audio Classification
|
16 |
+
type: audio-classification
|
17 |
+
dataset:
|
18 |
+
name: superb
|
19 |
+
type: superb
|
20 |
+
config: ks
|
21 |
+
split: validation
|
22 |
+
args: ks
|
23 |
+
metrics:
|
24 |
+
- name: Accuracy
|
25 |
+
type: accuracy
|
26 |
+
value: 0.9848484848484849
|
27 |
+
---
|
28 |
+
|
29 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
30 |
+
should probably proofread and complete it, then remove this comment. -->
|
31 |
+
|
32 |
+
# superb_ks_42
|
33 |
+
|
34 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the superb dataset.
|
35 |
+
It achieves the following results on the evaluation set:
|
36 |
+
- Loss: 0.0976
|
37 |
+
- Accuracy: 0.9848
|
38 |
+
|
39 |
+
## Model description
|
40 |
+
|
41 |
+
More information needed
|
42 |
+
|
43 |
+
## Intended uses & limitations
|
44 |
+
|
45 |
+
More information needed
|
46 |
+
|
47 |
+
## Training and evaluation data
|
48 |
+
|
49 |
+
More information needed
|
50 |
+
|
51 |
+
## Training procedure
|
52 |
+
|
53 |
+
### Training hyperparameters
|
54 |
+
|
55 |
+
The following hyperparameters were used during training:
|
56 |
+
- learning_rate: 5e-05
|
57 |
+
- train_batch_size: 32
|
58 |
+
- eval_batch_size: 4
|
59 |
+
- seed: 42
|
60 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
61 |
+
- lr_scheduler_type: linear
|
62 |
+
- lr_scheduler_warmup_ratio: 0.1
|
63 |
+
- num_epochs: 10
|
64 |
+
|
65 |
+
### Training results
|
66 |
+
|
67 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
68 |
+
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
|
69 |
+
| 0.3599 | 1.0 | 1597 | 0.1546 | 0.9707 |
|
70 |
+
| 0.0819 | 2.0 | 3194 | 0.0998 | 0.9762 |
|
71 |
+
| 0.0635 | 3.0 | 4791 | 0.1049 | 0.9800 |
|
72 |
+
| 0.0437 | 4.0 | 6388 | 0.0905 | 0.9797 |
|
73 |
+
| 0.0411 | 5.0 | 7985 | 0.0898 | 0.9809 |
|
74 |
+
| 0.0283 | 6.0 | 9582 | 0.1006 | 0.9812 |
|
75 |
+
| 0.0229 | 7.0 | 11179 | 0.0976 | 0.9848 |
|
76 |
+
| 0.0186 | 8.0 | 12776 | 0.1143 | 0.9825 |
|
77 |
+
| 0.0094 | 9.0 | 14373 | 0.1136 | 0.9835 |
|
78 |
+
| 0.0066 | 10.0 | 15970 | 0.1172 | 0.9834 |
|
79 |
+
|
80 |
+
|
81 |
+
### Framework versions
|
82 |
+
|
83 |
+
- Transformers 4.40.1
|
84 |
+
- Pytorch 2.3.0+cu121
|
85 |
+
- Datasets 2.19.0
|
86 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"eval_accuracy": 0.9848484848484849,
|
4 |
+
"eval_loss": 0.09762121737003326,
|
5 |
+
"eval_runtime": 138.169,
|
6 |
+
"eval_samples_per_second": 49.201,
|
7 |
+
"eval_steps_per_second": 0.195,
|
8 |
+
"test_accuracy": 0.9383317104836092,
|
9 |
+
"test_loss": 0.5075491666793823,
|
10 |
+
"test_runtime": 68.0743,
|
11 |
+
"test_samples_per_second": 45.259,
|
12 |
+
"test_steps_per_second": 0.191,
|
13 |
+
"total_flos": 6.415922957241599e+19,
|
14 |
+
"train_accuracy": 0.9941676126355345,
|
15 |
+
"train_loss": 0.028768757358193398,
|
16 |
+
"train_runtime": 921.5231,
|
17 |
+
"train_samples_per_second": 55.445,
|
18 |
+
"train_steps_per_second": 13.862
|
19 |
+
}
|
config.json
ADDED
@@ -0,0 +1,181 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "openai/whisper-small",
|
3 |
+
"activation_dropout": 0.0,
|
4 |
+
"activation_function": "gelu",
|
5 |
+
"apply_spec_augment": false,
|
6 |
+
"architectures": [
|
7 |
+
"WhisperForAudioClassification"
|
8 |
+
],
|
9 |
+
"attention_dropout": 0.0,
|
10 |
+
"begin_suppress_tokens": [
|
11 |
+
220,
|
12 |
+
50257
|
13 |
+
],
|
14 |
+
"bos_token_id": 50257,
|
15 |
+
"classifier_proj_size": 256,
|
16 |
+
"d_model": 768,
|
17 |
+
"decoder_attention_heads": 12,
|
18 |
+
"decoder_ffn_dim": 3072,
|
19 |
+
"decoder_layerdrop": 0.0,
|
20 |
+
"decoder_layers": 12,
|
21 |
+
"decoder_start_token_id": 50258,
|
22 |
+
"dropout": 0.0,
|
23 |
+
"encoder_attention_heads": 12,
|
24 |
+
"encoder_ffn_dim": 3072,
|
25 |
+
"encoder_layerdrop": 0.0,
|
26 |
+
"encoder_layers": 12,
|
27 |
+
"eos_token_id": 50257,
|
28 |
+
"finetuning_task": "audio-classification",
|
29 |
+
"forced_decoder_ids": [
|
30 |
+
[
|
31 |
+
1,
|
32 |
+
50259
|
33 |
+
],
|
34 |
+
[
|
35 |
+
2,
|
36 |
+
50359
|
37 |
+
],
|
38 |
+
[
|
39 |
+
3,
|
40 |
+
50363
|
41 |
+
]
|
42 |
+
],
|
43 |
+
"id2label": {
|
44 |
+
"0": "yes",
|
45 |
+
"1": "no",
|
46 |
+
"10": "_silence_",
|
47 |
+
"11": "_unknown_",
|
48 |
+
"2": "up",
|
49 |
+
"3": "down",
|
50 |
+
"4": "left",
|
51 |
+
"5": "right",
|
52 |
+
"6": "on",
|
53 |
+
"7": "off",
|
54 |
+
"8": "stop",
|
55 |
+
"9": "go"
|
56 |
+
},
|
57 |
+
"init_std": 0.02,
|
58 |
+
"is_encoder_decoder": true,
|
59 |
+
"label2id": {
|
60 |
+
"_silence_": "10",
|
61 |
+
"_unknown_": "11",
|
62 |
+
"down": "3",
|
63 |
+
"go": "9",
|
64 |
+
"left": "4",
|
65 |
+
"no": "1",
|
66 |
+
"off": "7",
|
67 |
+
"on": "6",
|
68 |
+
"right": "5",
|
69 |
+
"stop": "8",
|
70 |
+
"up": "2",
|
71 |
+
"yes": "0"
|
72 |
+
},
|
73 |
+
"mask_feature_length": 10,
|
74 |
+
"mask_feature_min_masks": 0,
|
75 |
+
"mask_feature_prob": 0.0,
|
76 |
+
"mask_time_length": 10,
|
77 |
+
"mask_time_min_masks": 2,
|
78 |
+
"mask_time_prob": 0.05,
|
79 |
+
"max_length": 448,
|
80 |
+
"max_source_positions": 1500,
|
81 |
+
"max_target_positions": 448,
|
82 |
+
"median_filter_width": 7,
|
83 |
+
"model_type": "whisper",
|
84 |
+
"num_hidden_layers": 12,
|
85 |
+
"num_mel_bins": 80,
|
86 |
+
"pad_token_id": 50257,
|
87 |
+
"scale_embedding": false,
|
88 |
+
"suppress_tokens": [
|
89 |
+
1,
|
90 |
+
2,
|
91 |
+
7,
|
92 |
+
8,
|
93 |
+
9,
|
94 |
+
10,
|
95 |
+
14,
|
96 |
+
25,
|
97 |
+
26,
|
98 |
+
27,
|
99 |
+
28,
|
100 |
+
29,
|
101 |
+
31,
|
102 |
+
58,
|
103 |
+
59,
|
104 |
+
60,
|
105 |
+
61,
|
106 |
+
62,
|
107 |
+
63,
|
108 |
+
90,
|
109 |
+
91,
|
110 |
+
92,
|
111 |
+
93,
|
112 |
+
359,
|
113 |
+
503,
|
114 |
+
522,
|
115 |
+
542,
|
116 |
+
873,
|
117 |
+
893,
|
118 |
+
902,
|
119 |
+
918,
|
120 |
+
922,
|
121 |
+
931,
|
122 |
+
1350,
|
123 |
+
1853,
|
124 |
+
1982,
|
125 |
+
2460,
|
126 |
+
2627,
|
127 |
+
3246,
|
128 |
+
3253,
|
129 |
+
3268,
|
130 |
+
3536,
|
131 |
+
3846,
|
132 |
+
3961,
|
133 |
+
4183,
|
134 |
+
4667,
|
135 |
+
6585,
|
136 |
+
6647,
|
137 |
+
7273,
|
138 |
+
9061,
|
139 |
+
9383,
|
140 |
+
10428,
|
141 |
+
10929,
|
142 |
+
11938,
|
143 |
+
12033,
|
144 |
+
12331,
|
145 |
+
12562,
|
146 |
+
13793,
|
147 |
+
14157,
|
148 |
+
14635,
|
149 |
+
15265,
|
150 |
+
15618,
|
151 |
+
16553,
|
152 |
+
16604,
|
153 |
+
18362,
|
154 |
+
18956,
|
155 |
+
20075,
|
156 |
+
21675,
|
157 |
+
22520,
|
158 |
+
26130,
|
159 |
+
26161,
|
160 |
+
26435,
|
161 |
+
28279,
|
162 |
+
29464,
|
163 |
+
31650,
|
164 |
+
32302,
|
165 |
+
32470,
|
166 |
+
36865,
|
167 |
+
42863,
|
168 |
+
47425,
|
169 |
+
49870,
|
170 |
+
50254,
|
171 |
+
50258,
|
172 |
+
50360,
|
173 |
+
50361,
|
174 |
+
50362
|
175 |
+
],
|
176 |
+
"torch_dtype": "float32",
|
177 |
+
"transformers_version": "4.40.1",
|
178 |
+
"use_cache": true,
|
179 |
+
"use_weighted_layer_sum": false,
|
180 |
+
"vocab_size": 51865
|
181 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"eval_accuracy": 0.9848484848484849,
|
4 |
+
"eval_loss": 0.09762121737003326,
|
5 |
+
"eval_runtime": 138.169,
|
6 |
+
"eval_samples_per_second": 49.201,
|
7 |
+
"eval_steps_per_second": 0.195
|
8 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fdb1b8c84b7e2ffd638317a96698440612045be22bd2e7663a3c2453125d4924
|
3 |
+
size 353436648
|
pred_logit_eval.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:83155745ad5509e3078c211305b8f9c9f6af7158260174b6bd4526a2e5dcf6b7
|
3 |
+
size 326432
|
pred_logit_test.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8818682aa32ec15f5f35e7d4adf2fb779a1e979aab03da0500138a0969c13052
|
3 |
+
size 148016
|
pred_logit_train.npy
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:09e92a7e03681287a1f2c593946bc21cd5c0e5fbed9d35d517884024b6c3c511
|
3 |
+
size 2452640
|
preprocessor_config.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"chunk_length": 30,
|
3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
4 |
+
"feature_size": 80,
|
5 |
+
"hop_length": 160,
|
6 |
+
"n_fft": 400,
|
7 |
+
"n_samples": 480000,
|
8 |
+
"nb_max_frames": 3000,
|
9 |
+
"padding_side": "right",
|
10 |
+
"padding_value": 0.0,
|
11 |
+
"processor_class": "WhisperProcessor",
|
12 |
+
"return_attention_mask": false,
|
13 |
+
"sampling_rate": 16000
|
14 |
+
}
|
test_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"test_accuracy": 0.9383317104836092,
|
4 |
+
"test_loss": 0.5075491666793823,
|
5 |
+
"test_runtime": 68.0743,
|
6 |
+
"test_samples_per_second": 45.259,
|
7 |
+
"test_steps_per_second": 0.191
|
8 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 10.0,
|
3 |
+
"train_accuracy": 0.9941676126355345,
|
4 |
+
"train_loss": 0.028768757358193398,
|
5 |
+
"train_runtime": 921.5231,
|
6 |
+
"train_samples_per_second": 55.445,
|
7 |
+
"train_steps_per_second": 13.862
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,225 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.9848484848484849,
|
3 |
+
"best_model_checkpoint": "audio/train/checkpoint/whisper-small/superb_ks_42/checkpoint-11179",
|
4 |
+
"epoch": 10.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 15970,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.6261740763932373,
|
13 |
+
"grad_norm": 8.361194610595703,
|
14 |
+
"learning_rate": 3.1308703819661866e-05,
|
15 |
+
"loss": 0.3599,
|
16 |
+
"step": 1000
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 1.0,
|
20 |
+
"eval_accuracy": 0.9707266843189173,
|
21 |
+
"eval_loss": 0.15458986163139343,
|
22 |
+
"eval_runtime": 143.1918,
|
23 |
+
"eval_samples_per_second": 47.475,
|
24 |
+
"eval_steps_per_second": 0.189,
|
25 |
+
"step": 1597
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 1.2523481527864746,
|
29 |
+
"grad_norm": 0.06838619709014893,
|
30 |
+
"learning_rate": 4.859806581785292e-05,
|
31 |
+
"loss": 0.0958,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.878522229179712,
|
36 |
+
"grad_norm": 6.442350387573242,
|
37 |
+
"learning_rate": 4.51193209490016e-05,
|
38 |
+
"loss": 0.0819,
|
39 |
+
"step": 3000
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 2.0,
|
43 |
+
"eval_accuracy": 0.9761694616063548,
|
44 |
+
"eval_loss": 0.09978443384170532,
|
45 |
+
"eval_runtime": 138.5727,
|
46 |
+
"eval_samples_per_second": 49.057,
|
47 |
+
"eval_steps_per_second": 0.195,
|
48 |
+
"step": 3194
|
49 |
+
},
|
50 |
+
{
|
51 |
+
"epoch": 2.504696305572949,
|
52 |
+
"grad_norm": 0.008366833440959454,
|
53 |
+
"learning_rate": 4.1640576080150286e-05,
|
54 |
+
"loss": 0.0635,
|
55 |
+
"step": 4000
|
56 |
+
},
|
57 |
+
{
|
58 |
+
"epoch": 3.0,
|
59 |
+
"eval_accuracy": 0.979994115916446,
|
60 |
+
"eval_loss": 0.10489170253276825,
|
61 |
+
"eval_runtime": 138.2853,
|
62 |
+
"eval_samples_per_second": 49.159,
|
63 |
+
"eval_steps_per_second": 0.195,
|
64 |
+
"step": 4791
|
65 |
+
},
|
66 |
+
{
|
67 |
+
"epoch": 3.1308703819661865,
|
68 |
+
"grad_norm": 0.06020672991871834,
|
69 |
+
"learning_rate": 3.8161831211298965e-05,
|
70 |
+
"loss": 0.0569,
|
71 |
+
"step": 5000
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 3.757044458359424,
|
75 |
+
"grad_norm": 0.005801007151603699,
|
76 |
+
"learning_rate": 3.468308634244764e-05,
|
77 |
+
"loss": 0.0437,
|
78 |
+
"step": 6000
|
79 |
+
},
|
80 |
+
{
|
81 |
+
"epoch": 4.0,
|
82 |
+
"eval_accuracy": 0.9796999117387467,
|
83 |
+
"eval_loss": 0.0905165895819664,
|
84 |
+
"eval_runtime": 138.7617,
|
85 |
+
"eval_samples_per_second": 48.99,
|
86 |
+
"eval_steps_per_second": 0.195,
|
87 |
+
"step": 6388
|
88 |
+
},
|
89 |
+
{
|
90 |
+
"epoch": 4.383218534752661,
|
91 |
+
"grad_norm": 1.9717774391174316,
|
92 |
+
"learning_rate": 3.120434147359633e-05,
|
93 |
+
"loss": 0.0411,
|
94 |
+
"step": 7000
|
95 |
+
},
|
96 |
+
{
|
97 |
+
"epoch": 5.0,
|
98 |
+
"eval_accuracy": 0.980876728449544,
|
99 |
+
"eval_loss": 0.08979170769453049,
|
100 |
+
"eval_runtime": 138.2656,
|
101 |
+
"eval_samples_per_second": 49.166,
|
102 |
+
"eval_steps_per_second": 0.195,
|
103 |
+
"step": 7985
|
104 |
+
},
|
105 |
+
{
|
106 |
+
"epoch": 5.009392611145898,
|
107 |
+
"grad_norm": 0.09022530913352966,
|
108 |
+
"learning_rate": 2.7725596604745007e-05,
|
109 |
+
"loss": 0.0386,
|
110 |
+
"step": 8000
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 5.635566687539136,
|
114 |
+
"grad_norm": 0.003507192712277174,
|
115 |
+
"learning_rate": 2.4246851735893692e-05,
|
116 |
+
"loss": 0.0283,
|
117 |
+
"step": 9000
|
118 |
+
},
|
119 |
+
{
|
120 |
+
"epoch": 6.0,
|
121 |
+
"eval_accuracy": 0.9811709326272433,
|
122 |
+
"eval_loss": 0.10055801272392273,
|
123 |
+
"eval_runtime": 139.0437,
|
124 |
+
"eval_samples_per_second": 48.891,
|
125 |
+
"eval_steps_per_second": 0.194,
|
126 |
+
"step": 9582
|
127 |
+
},
|
128 |
+
{
|
129 |
+
"epoch": 6.261740763932373,
|
130 |
+
"grad_norm": 0.6481895446777344,
|
131 |
+
"learning_rate": 2.076810686704237e-05,
|
132 |
+
"loss": 0.0322,
|
133 |
+
"step": 10000
|
134 |
+
},
|
135 |
+
{
|
136 |
+
"epoch": 6.88791484032561,
|
137 |
+
"grad_norm": 15.71690559387207,
|
138 |
+
"learning_rate": 1.7289361998191053e-05,
|
139 |
+
"loss": 0.0229,
|
140 |
+
"step": 11000
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 7.0,
|
144 |
+
"eval_accuracy": 0.9848484848484849,
|
145 |
+
"eval_loss": 0.09762121737003326,
|
146 |
+
"eval_runtime": 138.2722,
|
147 |
+
"eval_samples_per_second": 49.164,
|
148 |
+
"eval_steps_per_second": 0.195,
|
149 |
+
"step": 11179
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 7.514088916718848,
|
153 |
+
"grad_norm": 0.0020187441259622574,
|
154 |
+
"learning_rate": 1.3810617129339735e-05,
|
155 |
+
"loss": 0.0186,
|
156 |
+
"step": 12000
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 8.0,
|
160 |
+
"eval_accuracy": 0.9824948514268903,
|
161 |
+
"eval_loss": 0.11428509652614594,
|
162 |
+
"eval_runtime": 139.2486,
|
163 |
+
"eval_samples_per_second": 48.819,
|
164 |
+
"eval_steps_per_second": 0.194,
|
165 |
+
"step": 12776
|
166 |
+
},
|
167 |
+
{
|
168 |
+
"epoch": 8.140262993112085,
|
169 |
+
"grad_norm": 0.6445852518081665,
|
170 |
+
"learning_rate": 1.0331872260488417e-05,
|
171 |
+
"loss": 0.0159,
|
172 |
+
"step": 13000
|
173 |
+
},
|
174 |
+
{
|
175 |
+
"epoch": 8.766437069505322,
|
176 |
+
"grad_norm": 0.003636428155004978,
|
177 |
+
"learning_rate": 6.853127391637097e-06,
|
178 |
+
"loss": 0.0094,
|
179 |
+
"step": 14000
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 9.0,
|
183 |
+
"eval_accuracy": 0.9835245660488379,
|
184 |
+
"eval_loss": 0.11361410468816757,
|
185 |
+
"eval_runtime": 139.9155,
|
186 |
+
"eval_samples_per_second": 48.586,
|
187 |
+
"eval_steps_per_second": 0.193,
|
188 |
+
"step": 14373
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 9.39261114589856,
|
192 |
+
"grad_norm": 0.0017493839841336012,
|
193 |
+
"learning_rate": 3.3743825227857787e-06,
|
194 |
+
"loss": 0.0066,
|
195 |
+
"step": 15000
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 10.0,
|
199 |
+
"eval_accuracy": 0.9833774639599883,
|
200 |
+
"eval_loss": 0.11724763363599777,
|
201 |
+
"eval_runtime": 138.7939,
|
202 |
+
"eval_samples_per_second": 48.979,
|
203 |
+
"eval_steps_per_second": 0.195,
|
204 |
+
"step": 15970
|
205 |
+
},
|
206 |
+
{
|
207 |
+
"epoch": 10.0,
|
208 |
+
"step": 15970,
|
209 |
+
"total_flos": 6.415922957241599e+19,
|
210 |
+
"train_loss": 0.057567716346506634,
|
211 |
+
"train_runtime": 27132.3161,
|
212 |
+
"train_samples_per_second": 18.831,
|
213 |
+
"train_steps_per_second": 0.589
|
214 |
+
}
|
215 |
+
],
|
216 |
+
"logging_steps": 1000,
|
217 |
+
"max_steps": 15970,
|
218 |
+
"num_input_tokens_seen": 0,
|
219 |
+
"num_train_epochs": 10,
|
220 |
+
"save_steps": 500,
|
221 |
+
"total_flos": 6.415922957241599e+19,
|
222 |
+
"train_batch_size": 32,
|
223 |
+
"trial_name": null,
|
224 |
+
"trial_params": null
|
225 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:08f07a729079763eab0fd2151437a9f3c4e6fbf2c0bee81734b59d0d4c00efba
|
3 |
+
size 5112
|