End of training
Browse files- README.md +16 -4
- all_results.json +15 -0
- eval_results.json +9 -0
- runs/Jan06_18-30-52_srvrocgpu011.uct.ac.za/events.out.tfevents.1736186470.srvrocgpu011.uct.ac.za +3 -0
- train_results.json +9 -0
- trainer_state.json +441 -0
README.md
CHANGED
@@ -4,11 +4,23 @@ license: apache-2.0
|
|
4 |
base_model: openai/whisper-large-v3
|
5 |
tags:
|
6 |
- generated_from_trainer
|
|
|
|
|
7 |
metrics:
|
8 |
- wer
|
9 |
model-index:
|
10 |
- name: whisper-large-v3-nyagen-balanced-model
|
11 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
---
|
13 |
|
14 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -16,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# whisper-large-v3-nyagen-balanced-model
|
18 |
|
19 |
-
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on
|
20 |
It achieves the following results on the evaluation set:
|
21 |
-
- Loss: 0.
|
22 |
-
- Wer: 0.
|
23 |
|
24 |
## Model description
|
25 |
|
|
|
4 |
base_model: openai/whisper-large-v3
|
5 |
tags:
|
6 |
- generated_from_trainer
|
7 |
+
datasets:
|
8 |
+
- nyagen
|
9 |
metrics:
|
10 |
- wer
|
11 |
model-index:
|
12 |
- name: whisper-large-v3-nyagen-balanced-model
|
13 |
+
results:
|
14 |
+
- task:
|
15 |
+
name: Automatic Speech Recognition
|
16 |
+
type: automatic-speech-recognition
|
17 |
+
dataset:
|
18 |
+
name: nyagen
|
19 |
+
type: nyagen
|
20 |
+
metrics:
|
21 |
+
- name: Wer
|
22 |
+
type: wer
|
23 |
+
value: 0.24026512013256007
|
24 |
---
|
25 |
|
26 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
28 |
|
29 |
# whisper-large-v3-nyagen-balanced-model
|
30 |
|
31 |
+
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the nyagen dataset.
|
32 |
It achieves the following results on the evaluation set:
|
33 |
+
- Loss: 0.3155
|
34 |
+
- Wer: 0.2403
|
35 |
|
36 |
## Model description
|
37 |
|
all_results.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 6.4534412955465585,
|
3 |
+
"eval_loss": 0.31553423404693604,
|
4 |
+
"eval_runtime": 178.7847,
|
5 |
+
"eval_samples": 177,
|
6 |
+
"eval_samples_per_second": 0.99,
|
7 |
+
"eval_steps_per_second": 0.498,
|
8 |
+
"eval_wer": 0.24026512013256007,
|
9 |
+
"total_flos": 3.247328881410048e+19,
|
10 |
+
"train_loss": 1.0919783552487692,
|
11 |
+
"train_runtime": 4827.0467,
|
12 |
+
"train_samples": 1481,
|
13 |
+
"train_samples_per_second": 8.287,
|
14 |
+
"train_steps_per_second": 1.036
|
15 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 6.4534412955465585,
|
3 |
+
"eval_loss": 0.31553423404693604,
|
4 |
+
"eval_runtime": 178.7847,
|
5 |
+
"eval_samples": 177,
|
6 |
+
"eval_samples_per_second": 0.99,
|
7 |
+
"eval_steps_per_second": 0.498,
|
8 |
+
"eval_wer": 0.24026512013256007
|
9 |
+
}
|
runs/Jan06_18-30-52_srvrocgpu011.uct.ac.za/events.out.tfevents.1736186470.srvrocgpu011.uct.ac.za
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:04cb78e1c8eb5e05fbad3ef9de53b13250436a32af6c469be34b6496f6cade75
|
3 |
+
size 40
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 6.4534412955465585,
|
3 |
+
"total_flos": 3.247328881410048e+19,
|
4 |
+
"train_loss": 1.0919783552487692,
|
5 |
+
"train_runtime": 4827.0467,
|
6 |
+
"train_samples": 1481,
|
7 |
+
"train_samples_per_second": 8.287,
|
8 |
+
"train_steps_per_second": 1.036
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,441 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.31553423404693604,
|
3 |
+
"best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-large-v3-nyagen-balanced-model/checkpoint-600",
|
4 |
+
"epoch": 6.4534412955465585,
|
5 |
+
"eval_steps": 200,
|
6 |
+
"global_step": 1200,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1349527665317139,
|
13 |
+
"grad_norm": 49.952491760253906,
|
14 |
+
"learning_rate": 4.6000000000000004e-07,
|
15 |
+
"loss": 9.6771,
|
16 |
+
"step": 25
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.2699055330634278,
|
20 |
+
"grad_norm": 29.110328674316406,
|
21 |
+
"learning_rate": 9.600000000000001e-07,
|
22 |
+
"loss": 6.8615,
|
23 |
+
"step": 50
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.4048582995951417,
|
27 |
+
"grad_norm": 32.85552215576172,
|
28 |
+
"learning_rate": 1.46e-06,
|
29 |
+
"loss": 4.6623,
|
30 |
+
"step": 75
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.5398110661268556,
|
34 |
+
"grad_norm": 27.510845184326172,
|
35 |
+
"learning_rate": 1.9600000000000003e-06,
|
36 |
+
"loss": 3.0651,
|
37 |
+
"step": 100
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.6747638326585695,
|
41 |
+
"grad_norm": 19.992250442504883,
|
42 |
+
"learning_rate": 2.46e-06,
|
43 |
+
"loss": 2.5318,
|
44 |
+
"step": 125
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.8097165991902834,
|
48 |
+
"grad_norm": 24.59845542907715,
|
49 |
+
"learning_rate": 2.96e-06,
|
50 |
+
"loss": 1.9071,
|
51 |
+
"step": 150
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.9446693657219973,
|
55 |
+
"grad_norm": 21.07068634033203,
|
56 |
+
"learning_rate": 3.46e-06,
|
57 |
+
"loss": 1.7429,
|
58 |
+
"step": 175
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 1.0755735492577598,
|
62 |
+
"grad_norm": 19.914566040039062,
|
63 |
+
"learning_rate": 3.96e-06,
|
64 |
+
"loss": 1.426,
|
65 |
+
"step": 200
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 1.0755735492577598,
|
69 |
+
"eval_loss": 0.4107988476753235,
|
70 |
+
"eval_runtime": 175.2165,
|
71 |
+
"eval_samples_per_second": 1.01,
|
72 |
+
"eval_steps_per_second": 0.508,
|
73 |
+
"eval_wer": 0.30695940347970174,
|
74 |
+
"step": 200
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 1.2105263157894737,
|
78 |
+
"grad_norm": 21.64095115661621,
|
79 |
+
"learning_rate": 4.4600000000000005e-06,
|
80 |
+
"loss": 1.4253,
|
81 |
+
"step": 225
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 1.3454790823211875,
|
85 |
+
"grad_norm": 28.434995651245117,
|
86 |
+
"learning_rate": 4.960000000000001e-06,
|
87 |
+
"loss": 1.3828,
|
88 |
+
"step": 250
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 1.4804318488529016,
|
92 |
+
"grad_norm": 17.18960189819336,
|
93 |
+
"learning_rate": 5.460000000000001e-06,
|
94 |
+
"loss": 1.2158,
|
95 |
+
"step": 275
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 1.6153846153846154,
|
99 |
+
"grad_norm": 23.611936569213867,
|
100 |
+
"learning_rate": 5.9600000000000005e-06,
|
101 |
+
"loss": 1.2998,
|
102 |
+
"step": 300
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 1.7503373819163293,
|
106 |
+
"grad_norm": 15.088561058044434,
|
107 |
+
"learning_rate": 6.460000000000001e-06,
|
108 |
+
"loss": 1.223,
|
109 |
+
"step": 325
|
110 |
+
},
|
111 |
+
{
|
112 |
+
"epoch": 1.885290148448043,
|
113 |
+
"grad_norm": 24.57098388671875,
|
114 |
+
"learning_rate": 6.96e-06,
|
115 |
+
"loss": 1.3904,
|
116 |
+
"step": 350
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 2.016194331983806,
|
120 |
+
"grad_norm": 10.937520027160645,
|
121 |
+
"learning_rate": 7.4600000000000006e-06,
|
122 |
+
"loss": 1.0776,
|
123 |
+
"step": 375
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 2.1511470985155197,
|
127 |
+
"grad_norm": 15.670218467712402,
|
128 |
+
"learning_rate": 7.960000000000002e-06,
|
129 |
+
"loss": 0.6798,
|
130 |
+
"step": 400
|
131 |
+
},
|
132 |
+
{
|
133 |
+
"epoch": 2.1511470985155197,
|
134 |
+
"eval_loss": 0.3342593014240265,
|
135 |
+
"eval_runtime": 182.0049,
|
136 |
+
"eval_samples_per_second": 0.973,
|
137 |
+
"eval_steps_per_second": 0.489,
|
138 |
+
"eval_wer": 0.2866611433305717,
|
139 |
+
"step": 400
|
140 |
+
},
|
141 |
+
{
|
142 |
+
"epoch": 2.2860998650472335,
|
143 |
+
"grad_norm": 11.230082511901855,
|
144 |
+
"learning_rate": 8.46e-06,
|
145 |
+
"loss": 0.6412,
|
146 |
+
"step": 425
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 2.4210526315789473,
|
150 |
+
"grad_norm": 14.338324546813965,
|
151 |
+
"learning_rate": 8.96e-06,
|
152 |
+
"loss": 0.6681,
|
153 |
+
"step": 450
|
154 |
+
},
|
155 |
+
{
|
156 |
+
"epoch": 2.556005398110661,
|
157 |
+
"grad_norm": 10.439035415649414,
|
158 |
+
"learning_rate": 9.460000000000001e-06,
|
159 |
+
"loss": 0.7359,
|
160 |
+
"step": 475
|
161 |
+
},
|
162 |
+
{
|
163 |
+
"epoch": 2.690958164642375,
|
164 |
+
"grad_norm": 17.129701614379883,
|
165 |
+
"learning_rate": 9.960000000000001e-06,
|
166 |
+
"loss": 0.7664,
|
167 |
+
"step": 500
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 2.8259109311740893,
|
171 |
+
"grad_norm": 26.702089309692383,
|
172 |
+
"learning_rate": 9.94888888888889e-06,
|
173 |
+
"loss": 0.8653,
|
174 |
+
"step": 525
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 2.960863697705803,
|
178 |
+
"grad_norm": 10.853477478027344,
|
179 |
+
"learning_rate": 9.893333333333334e-06,
|
180 |
+
"loss": 0.8808,
|
181 |
+
"step": 550
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 3.0917678812415654,
|
185 |
+
"grad_norm": 10.325875282287598,
|
186 |
+
"learning_rate": 9.837777777777778e-06,
|
187 |
+
"loss": 0.5139,
|
188 |
+
"step": 575
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 3.2267206477732793,
|
192 |
+
"grad_norm": 8.529032707214355,
|
193 |
+
"learning_rate": 9.782222222222222e-06,
|
194 |
+
"loss": 0.3574,
|
195 |
+
"step": 600
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 3.2267206477732793,
|
199 |
+
"eval_loss": 0.31553423404693604,
|
200 |
+
"eval_runtime": 179.3559,
|
201 |
+
"eval_samples_per_second": 0.987,
|
202 |
+
"eval_steps_per_second": 0.496,
|
203 |
+
"eval_wer": 0.24026512013256007,
|
204 |
+
"step": 600
|
205 |
+
},
|
206 |
+
{
|
207 |
+
"epoch": 3.361673414304993,
|
208 |
+
"grad_norm": 10.741698265075684,
|
209 |
+
"learning_rate": 9.726666666666668e-06,
|
210 |
+
"loss": 0.4186,
|
211 |
+
"step": 625
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 3.4966261808367074,
|
215 |
+
"grad_norm": 11.486895561218262,
|
216 |
+
"learning_rate": 9.671111111111112e-06,
|
217 |
+
"loss": 0.424,
|
218 |
+
"step": 650
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 3.6315789473684212,
|
222 |
+
"grad_norm": 11.22132682800293,
|
223 |
+
"learning_rate": 9.615555555555558e-06,
|
224 |
+
"loss": 0.3928,
|
225 |
+
"step": 675
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 3.766531713900135,
|
229 |
+
"grad_norm": 5.59958553314209,
|
230 |
+
"learning_rate": 9.56e-06,
|
231 |
+
"loss": 0.412,
|
232 |
+
"step": 700
|
233 |
+
},
|
234 |
+
{
|
235 |
+
"epoch": 3.901484480431849,
|
236 |
+
"grad_norm": 12.270620346069336,
|
237 |
+
"learning_rate": 9.504444444444446e-06,
|
238 |
+
"loss": 0.4351,
|
239 |
+
"step": 725
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 4.032388663967612,
|
243 |
+
"grad_norm": 7.528313636779785,
|
244 |
+
"learning_rate": 9.44888888888889e-06,
|
245 |
+
"loss": 0.3451,
|
246 |
+
"step": 750
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 4.1673414304993255,
|
250 |
+
"grad_norm": 9.45571517944336,
|
251 |
+
"learning_rate": 9.393333333333334e-06,
|
252 |
+
"loss": 0.1869,
|
253 |
+
"step": 775
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 4.302294197031039,
|
257 |
+
"grad_norm": 7.190002918243408,
|
258 |
+
"learning_rate": 9.33777777777778e-06,
|
259 |
+
"loss": 0.2691,
|
260 |
+
"step": 800
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 4.302294197031039,
|
264 |
+
"eval_loss": 0.33649739623069763,
|
265 |
+
"eval_runtime": 174.6568,
|
266 |
+
"eval_samples_per_second": 1.013,
|
267 |
+
"eval_steps_per_second": 0.51,
|
268 |
+
"eval_wer": 0.21582435791217897,
|
269 |
+
"step": 800
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 4.437246963562753,
|
273 |
+
"grad_norm": 5.602508068084717,
|
274 |
+
"learning_rate": 9.282222222222222e-06,
|
275 |
+
"loss": 0.2733,
|
276 |
+
"step": 825
|
277 |
+
},
|
278 |
+
{
|
279 |
+
"epoch": 4.572199730094467,
|
280 |
+
"grad_norm": 8.376420021057129,
|
281 |
+
"learning_rate": 9.226666666666668e-06,
|
282 |
+
"loss": 0.245,
|
283 |
+
"step": 850
|
284 |
+
},
|
285 |
+
{
|
286 |
+
"epoch": 4.707152496626181,
|
287 |
+
"grad_norm": 8.70338249206543,
|
288 |
+
"learning_rate": 9.171111111111112e-06,
|
289 |
+
"loss": 0.2119,
|
290 |
+
"step": 875
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 4.842105263157895,
|
294 |
+
"grad_norm": 8.216238975524902,
|
295 |
+
"learning_rate": 9.115555555555556e-06,
|
296 |
+
"loss": 0.2399,
|
297 |
+
"step": 900
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 4.9770580296896085,
|
301 |
+
"grad_norm": 8.389748573303223,
|
302 |
+
"learning_rate": 9.060000000000001e-06,
|
303 |
+
"loss": 0.2326,
|
304 |
+
"step": 925
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 5.107962213225371,
|
308 |
+
"grad_norm": 9.63129711151123,
|
309 |
+
"learning_rate": 9.004444444444445e-06,
|
310 |
+
"loss": 0.1235,
|
311 |
+
"step": 950
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 5.242914979757085,
|
315 |
+
"grad_norm": 7.203446865081787,
|
316 |
+
"learning_rate": 8.94888888888889e-06,
|
317 |
+
"loss": 0.111,
|
318 |
+
"step": 975
|
319 |
+
},
|
320 |
+
{
|
321 |
+
"epoch": 5.377867746288799,
|
322 |
+
"grad_norm": 9.31028938293457,
|
323 |
+
"learning_rate": 8.893333333333333e-06,
|
324 |
+
"loss": 0.1851,
|
325 |
+
"step": 1000
|
326 |
+
},
|
327 |
+
{
|
328 |
+
"epoch": 5.377867746288799,
|
329 |
+
"eval_loss": 0.31591707468032837,
|
330 |
+
"eval_runtime": 183.6608,
|
331 |
+
"eval_samples_per_second": 0.964,
|
332 |
+
"eval_steps_per_second": 0.485,
|
333 |
+
"eval_wer": 0.2903893951946976,
|
334 |
+
"step": 1000
|
335 |
+
},
|
336 |
+
{
|
337 |
+
"epoch": 5.512820512820513,
|
338 |
+
"grad_norm": 8.492629051208496,
|
339 |
+
"learning_rate": 8.83777777777778e-06,
|
340 |
+
"loss": 0.1237,
|
341 |
+
"step": 1025
|
342 |
+
},
|
343 |
+
{
|
344 |
+
"epoch": 5.647773279352227,
|
345 |
+
"grad_norm": 2.383087158203125,
|
346 |
+
"learning_rate": 8.782222222222223e-06,
|
347 |
+
"loss": 0.1468,
|
348 |
+
"step": 1050
|
349 |
+
},
|
350 |
+
{
|
351 |
+
"epoch": 5.7827260458839405,
|
352 |
+
"grad_norm": 6.9894232749938965,
|
353 |
+
"learning_rate": 8.726666666666667e-06,
|
354 |
+
"loss": 0.1299,
|
355 |
+
"step": 1075
|
356 |
+
},
|
357 |
+
{
|
358 |
+
"epoch": 5.917678812415654,
|
359 |
+
"grad_norm": 10.51385498046875,
|
360 |
+
"learning_rate": 8.671111111111113e-06,
|
361 |
+
"loss": 0.1686,
|
362 |
+
"step": 1100
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 6.048582995951417,
|
366 |
+
"grad_norm": 6.33004903793335,
|
367 |
+
"learning_rate": 8.615555555555555e-06,
|
368 |
+
"loss": 0.0867,
|
369 |
+
"step": 1125
|
370 |
+
},
|
371 |
+
{
|
372 |
+
"epoch": 6.183535762483131,
|
373 |
+
"grad_norm": 7.478524208068848,
|
374 |
+
"learning_rate": 8.560000000000001e-06,
|
375 |
+
"loss": 0.0918,
|
376 |
+
"step": 1150
|
377 |
+
},
|
378 |
+
{
|
379 |
+
"epoch": 6.318488529014845,
|
380 |
+
"grad_norm": 3.3907110691070557,
|
381 |
+
"learning_rate": 8.504444444444445e-06,
|
382 |
+
"loss": 0.0923,
|
383 |
+
"step": 1175
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 6.4534412955465585,
|
387 |
+
"grad_norm": 3.2987215518951416,
|
388 |
+
"learning_rate": 8.448888888888889e-06,
|
389 |
+
"loss": 0.0715,
|
390 |
+
"step": 1200
|
391 |
+
},
|
392 |
+
{
|
393 |
+
"epoch": 6.4534412955465585,
|
394 |
+
"eval_loss": 0.3675512373447418,
|
395 |
+
"eval_runtime": 175.2377,
|
396 |
+
"eval_samples_per_second": 1.01,
|
397 |
+
"eval_steps_per_second": 0.508,
|
398 |
+
"eval_wer": 0.2083678541839271,
|
399 |
+
"step": 1200
|
400 |
+
},
|
401 |
+
{
|
402 |
+
"epoch": 6.4534412955465585,
|
403 |
+
"step": 1200,
|
404 |
+
"total_flos": 3.247328881410048e+19,
|
405 |
+
"train_loss": 1.0919783552487692,
|
406 |
+
"train_runtime": 4827.0467,
|
407 |
+
"train_samples_per_second": 8.287,
|
408 |
+
"train_steps_per_second": 1.036
|
409 |
+
}
|
410 |
+
],
|
411 |
+
"logging_steps": 25,
|
412 |
+
"max_steps": 5000,
|
413 |
+
"num_input_tokens_seen": 0,
|
414 |
+
"num_train_epochs": 28,
|
415 |
+
"save_steps": 200,
|
416 |
+
"stateful_callbacks": {
|
417 |
+
"EarlyStoppingCallback": {
|
418 |
+
"args": {
|
419 |
+
"early_stopping_patience": 3,
|
420 |
+
"early_stopping_threshold": 0.0
|
421 |
+
},
|
422 |
+
"attributes": {
|
423 |
+
"early_stopping_patience_counter": 3
|
424 |
+
}
|
425 |
+
},
|
426 |
+
"TrainerControl": {
|
427 |
+
"args": {
|
428 |
+
"should_epoch_stop": false,
|
429 |
+
"should_evaluate": false,
|
430 |
+
"should_log": false,
|
431 |
+
"should_save": true,
|
432 |
+
"should_training_stop": true
|
433 |
+
},
|
434 |
+
"attributes": {}
|
435 |
+
}
|
436 |
+
},
|
437 |
+
"total_flos": 3.247328881410048e+19,
|
438 |
+
"train_batch_size": 2,
|
439 |
+
"trial_name": null,
|
440 |
+
"trial_params": null
|
441 |
+
}
|