End of training
Browse files- README.md +7 -7
- model.safetensors +1 -1
- trainer_state.json +322 -0
README.md
CHANGED
@@ -21,13 +21,13 @@ should probably proofread and complete it, then remove this comment. -->
|
|
21 |
|
22 |
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the AshtonLKY/augmented_audio dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
-
- eval_loss: 0.
|
25 |
-
- eval_wer:
|
26 |
-
- eval_runtime:
|
27 |
-
- eval_samples_per_second: 2.
|
28 |
-
- eval_steps_per_second: 0.
|
29 |
-
- epoch:
|
30 |
-
- step:
|
31 |
|
32 |
## Model description
|
33 |
|
|
|
21 |
|
22 |
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the AshtonLKY/augmented_audio dataset.
|
23 |
It achieves the following results on the evaluation set:
|
24 |
+
- eval_loss: 0.0918
|
25 |
+
- eval_wer: 6.4944
|
26 |
+
- eval_runtime: 5293.1212
|
27 |
+
- eval_samples_per_second: 2.539
|
28 |
+
- eval_steps_per_second: 0.159
|
29 |
+
- epoch: 0.5952
|
30 |
+
- step: 2000
|
31 |
|
32 |
## Model description
|
33 |
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 966995080
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:917229462f303857f3cc2852914e3560bb1370b32e4f34a44f6dd649f9ca7326
|
3 |
size 966995080
|
trainer_state.json
ADDED
@@ -0,0 +1,322 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 6.12266119021778,
|
3 |
+
"best_model_checkpoint": "AshtonLKY/Whisper_ASR_ATC_v6/checkpoint-1000",
|
4 |
+
"epoch": 0.2976190476190476,
|
5 |
+
"eval_steps": 1000,
|
6 |
+
"global_step": 1000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.00744047619047619,
|
13 |
+
"grad_norm": 5.413925647735596,
|
14 |
+
"learning_rate": 5.000000000000001e-07,
|
15 |
+
"loss": 0.0412,
|
16 |
+
"step": 25
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.01488095238095238,
|
20 |
+
"grad_norm": 3.476644515991211,
|
21 |
+
"learning_rate": 1.0000000000000002e-06,
|
22 |
+
"loss": 0.0409,
|
23 |
+
"step": 50
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.022321428571428572,
|
27 |
+
"grad_norm": 1.9877393245697021,
|
28 |
+
"learning_rate": 1.5e-06,
|
29 |
+
"loss": 0.0357,
|
30 |
+
"step": 75
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.02976190476190476,
|
34 |
+
"grad_norm": 1.821871280670166,
|
35 |
+
"learning_rate": 2.0000000000000003e-06,
|
36 |
+
"loss": 0.0297,
|
37 |
+
"step": 100
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.03720238095238095,
|
41 |
+
"grad_norm": 4.649389743804932,
|
42 |
+
"learning_rate": 2.5e-06,
|
43 |
+
"loss": 0.0284,
|
44 |
+
"step": 125
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.044642857142857144,
|
48 |
+
"grad_norm": 3.751368999481201,
|
49 |
+
"learning_rate": 3e-06,
|
50 |
+
"loss": 0.026,
|
51 |
+
"step": 150
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.052083333333333336,
|
55 |
+
"grad_norm": 5.829326152801514,
|
56 |
+
"learning_rate": 3.5e-06,
|
57 |
+
"loss": 0.0278,
|
58 |
+
"step": 175
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.05952380952380952,
|
62 |
+
"grad_norm": 2.7991154193878174,
|
63 |
+
"learning_rate": 4.000000000000001e-06,
|
64 |
+
"loss": 0.0215,
|
65 |
+
"step": 200
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.06696428571428571,
|
69 |
+
"grad_norm": 1.027055025100708,
|
70 |
+
"learning_rate": 4.5e-06,
|
71 |
+
"loss": 0.0131,
|
72 |
+
"step": 225
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.0744047619047619,
|
76 |
+
"grad_norm": 2.6203041076660156,
|
77 |
+
"learning_rate": 5e-06,
|
78 |
+
"loss": 0.0237,
|
79 |
+
"step": 250
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.0818452380952381,
|
83 |
+
"grad_norm": 0.10589638352394104,
|
84 |
+
"learning_rate": 5.500000000000001e-06,
|
85 |
+
"loss": 0.0154,
|
86 |
+
"step": 275
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.08928571428571429,
|
90 |
+
"grad_norm": 0.6042885184288025,
|
91 |
+
"learning_rate": 6e-06,
|
92 |
+
"loss": 0.0138,
|
93 |
+
"step": 300
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.09672619047619048,
|
97 |
+
"grad_norm": 4.635740280151367,
|
98 |
+
"learning_rate": 6.5000000000000004e-06,
|
99 |
+
"loss": 0.012,
|
100 |
+
"step": 325
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.10416666666666667,
|
104 |
+
"grad_norm": 1.3288782835006714,
|
105 |
+
"learning_rate": 7e-06,
|
106 |
+
"loss": 0.0105,
|
107 |
+
"step": 350
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.11160714285714286,
|
111 |
+
"grad_norm": 0.3894687592983246,
|
112 |
+
"learning_rate": 7.500000000000001e-06,
|
113 |
+
"loss": 0.01,
|
114 |
+
"step": 375
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.11904761904761904,
|
118 |
+
"grad_norm": 2.7233240604400635,
|
119 |
+
"learning_rate": 8.000000000000001e-06,
|
120 |
+
"loss": 0.0104,
|
121 |
+
"step": 400
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.12648809523809523,
|
125 |
+
"grad_norm": 1.8257516622543335,
|
126 |
+
"learning_rate": 8.5e-06,
|
127 |
+
"loss": 0.0102,
|
128 |
+
"step": 425
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.13392857142857142,
|
132 |
+
"grad_norm": 20.001121520996094,
|
133 |
+
"learning_rate": 9e-06,
|
134 |
+
"loss": 0.0135,
|
135 |
+
"step": 450
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.14136904761904762,
|
139 |
+
"grad_norm": 1.909671425819397,
|
140 |
+
"learning_rate": 9.5e-06,
|
141 |
+
"loss": 0.0058,
|
142 |
+
"step": 475
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.1488095238095238,
|
146 |
+
"grad_norm": 0.6188738942146301,
|
147 |
+
"learning_rate": 1e-05,
|
148 |
+
"loss": 0.0108,
|
149 |
+
"step": 500
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.15625,
|
153 |
+
"grad_norm": 4.2201385498046875,
|
154 |
+
"learning_rate": 9.989795918367348e-06,
|
155 |
+
"loss": 0.01,
|
156 |
+
"step": 525
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.1636904761904762,
|
160 |
+
"grad_norm": 0.09348121285438538,
|
161 |
+
"learning_rate": 9.979591836734694e-06,
|
162 |
+
"loss": 0.0072,
|
163 |
+
"step": 550
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.17113095238095238,
|
167 |
+
"grad_norm": 0.8602906465530396,
|
168 |
+
"learning_rate": 9.969387755102042e-06,
|
169 |
+
"loss": 0.0074,
|
170 |
+
"step": 575
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.17857142857142858,
|
174 |
+
"grad_norm": 0.9500374794006348,
|
175 |
+
"learning_rate": 9.959183673469387e-06,
|
176 |
+
"loss": 0.01,
|
177 |
+
"step": 600
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.18601190476190477,
|
181 |
+
"grad_norm": 7.845817565917969,
|
182 |
+
"learning_rate": 9.948979591836737e-06,
|
183 |
+
"loss": 0.0158,
|
184 |
+
"step": 625
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.19345238095238096,
|
188 |
+
"grad_norm": 0.8142985701560974,
|
189 |
+
"learning_rate": 9.938775510204082e-06,
|
190 |
+
"loss": 0.0098,
|
191 |
+
"step": 650
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.20089285714285715,
|
195 |
+
"grad_norm": 4.271408557891846,
|
196 |
+
"learning_rate": 9.92857142857143e-06,
|
197 |
+
"loss": 0.009,
|
198 |
+
"step": 675
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.20833333333333334,
|
202 |
+
"grad_norm": 5.065946102142334,
|
203 |
+
"learning_rate": 9.918367346938776e-06,
|
204 |
+
"loss": 0.0155,
|
205 |
+
"step": 700
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.21577380952380953,
|
209 |
+
"grad_norm": 4.185829162597656,
|
210 |
+
"learning_rate": 9.908163265306123e-06,
|
211 |
+
"loss": 0.0083,
|
212 |
+
"step": 725
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.22321428571428573,
|
216 |
+
"grad_norm": 0.18144944310188293,
|
217 |
+
"learning_rate": 9.89795918367347e-06,
|
218 |
+
"loss": 0.0095,
|
219 |
+
"step": 750
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.23065476190476192,
|
223 |
+
"grad_norm": 1.0089025497436523,
|
224 |
+
"learning_rate": 9.887755102040816e-06,
|
225 |
+
"loss": 0.0149,
|
226 |
+
"step": 775
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.23809523809523808,
|
230 |
+
"grad_norm": 0.15361322462558746,
|
231 |
+
"learning_rate": 9.877551020408164e-06,
|
232 |
+
"loss": 0.0107,
|
233 |
+
"step": 800
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.24553571428571427,
|
237 |
+
"grad_norm": 0.5313578248023987,
|
238 |
+
"learning_rate": 9.867346938775511e-06,
|
239 |
+
"loss": 0.0151,
|
240 |
+
"step": 825
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.25297619047619047,
|
244 |
+
"grad_norm": 0.6949892044067383,
|
245 |
+
"learning_rate": 9.857142857142859e-06,
|
246 |
+
"loss": 0.0112,
|
247 |
+
"step": 850
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.2604166666666667,
|
251 |
+
"grad_norm": 3.9174296855926514,
|
252 |
+
"learning_rate": 9.846938775510205e-06,
|
253 |
+
"loss": 0.0108,
|
254 |
+
"step": 875
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.26785714285714285,
|
258 |
+
"grad_norm": 3.3171513080596924,
|
259 |
+
"learning_rate": 9.836734693877552e-06,
|
260 |
+
"loss": 0.0106,
|
261 |
+
"step": 900
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.27529761904761907,
|
265 |
+
"grad_norm": 1.71816086769104,
|
266 |
+
"learning_rate": 9.8265306122449e-06,
|
267 |
+
"loss": 0.0144,
|
268 |
+
"step": 925
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.28273809523809523,
|
272 |
+
"grad_norm": 2.81546688079834,
|
273 |
+
"learning_rate": 9.816326530612245e-06,
|
274 |
+
"loss": 0.009,
|
275 |
+
"step": 950
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.29017857142857145,
|
279 |
+
"grad_norm": 5.431806564331055,
|
280 |
+
"learning_rate": 9.806122448979593e-06,
|
281 |
+
"loss": 0.022,
|
282 |
+
"step": 975
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.2976190476190476,
|
286 |
+
"grad_norm": 4.169656753540039,
|
287 |
+
"learning_rate": 9.795918367346939e-06,
|
288 |
+
"loss": 0.0125,
|
289 |
+
"step": 1000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.2976190476190476,
|
293 |
+
"eval_loss": 0.09145254641771317,
|
294 |
+
"eval_runtime": 5258.4593,
|
295 |
+
"eval_samples_per_second": 2.556,
|
296 |
+
"eval_steps_per_second": 0.16,
|
297 |
+
"eval_wer": 6.12266119021778,
|
298 |
+
"step": 1000
|
299 |
+
}
|
300 |
+
],
|
301 |
+
"logging_steps": 25,
|
302 |
+
"max_steps": 25000,
|
303 |
+
"num_input_tokens_seen": 0,
|
304 |
+
"num_train_epochs": 8,
|
305 |
+
"save_steps": 1000,
|
306 |
+
"stateful_callbacks": {
|
307 |
+
"TrainerControl": {
|
308 |
+
"args": {
|
309 |
+
"should_epoch_stop": false,
|
310 |
+
"should_evaluate": false,
|
311 |
+
"should_log": false,
|
312 |
+
"should_save": true,
|
313 |
+
"should_training_stop": false
|
314 |
+
},
|
315 |
+
"attributes": {}
|
316 |
+
}
|
317 |
+
},
|
318 |
+
"total_flos": 4.61736640512e+18,
|
319 |
+
"train_batch_size": 16,
|
320 |
+
"trial_name": null,
|
321 |
+
"trial_params": null
|
322 |
+
}
|