Commit
•
d086623
1
Parent(s):
74b237f
End of training
Browse files- README.md +3 -1
- all_results.json +11 -11
- eval_results.json +6 -6
- runs/Aug30_21-53-10_4b35055fdbcb/events.out.tfevents.1725061483.4b35055fdbcb.318.1 +3 -0
- train_results.json +6 -6
- trainer_state.json +234 -106
README.md
CHANGED
@@ -2,6 +2,8 @@
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/wav2vec2-large-xlsr-53
|
4 |
tags:
|
|
|
|
|
5 |
- generated_from_trainer
|
6 |
metrics:
|
7 |
- wer
|
@@ -15,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# wav2vec2-xlsr-53-ft-btb-ccv-cy
|
17 |
|
18 |
-
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on
|
19 |
It achieves the following results on the evaluation set:
|
20 |
- Loss: 0.5226
|
21 |
- Wer: 0.4001
|
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/wav2vec2-large-xlsr-53
|
4 |
tags:
|
5 |
+
- automatic-speech-recognition
|
6 |
+
- DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
|
7 |
- generated_from_trainer
|
8 |
metrics:
|
9 |
- wer
|
|
|
17 |
|
18 |
# wav2vec2-xlsr-53-ft-btb-ccv-cy
|
19 |
|
20 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN-WITH-CCV - DEFAULT dataset.
|
21 |
It achieves the following results on the evaluation set:
|
22 |
- Loss: 0.5226
|
23 |
- Wer: 0.4001
|
all_results.json
CHANGED
@@ -1,15 +1,15 @@
|
|
1 |
{
|
2 |
-
"epoch": 0.
|
3 |
-
"eval_loss": 0.
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 7022,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second":
|
8 |
-
"eval_wer": 0.
|
9 |
-
"total_flos":
|
10 |
-
"train_loss": 0.
|
11 |
-
"train_runtime":
|
12 |
"train_samples": 124748,
|
13 |
-
"train_samples_per_second": 10.
|
14 |
-
"train_steps_per_second": 1.
|
15 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 0.6412722842118763,
|
3 |
+
"eval_loss": 0.5226185917854309,
|
4 |
+
"eval_runtime": 188.9199,
|
5 |
"eval_samples": 7022,
|
6 |
+
"eval_samples_per_second": 37.169,
|
7 |
+
"eval_steps_per_second": 0.582,
|
8 |
+
"eval_wer": 0.40009832711037885,
|
9 |
+
"total_flos": 1.1393778193380235e+19,
|
10 |
+
"train_loss": 0.7283544036865235,
|
11 |
+
"train_runtime": 7737.7643,
|
12 |
"train_samples": 124748,
|
13 |
+
"train_samples_per_second": 10.339,
|
14 |
+
"train_steps_per_second": 1.292
|
15 |
}
|
eval_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"epoch": 0.
|
3 |
-
"eval_loss": 0.
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 7022,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second":
|
8 |
-
"eval_wer": 0.
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 0.6412722842118763,
|
3 |
+
"eval_loss": 0.5226185917854309,
|
4 |
+
"eval_runtime": 188.9199,
|
5 |
"eval_samples": 7022,
|
6 |
+
"eval_samples_per_second": 37.169,
|
7 |
+
"eval_steps_per_second": 0.582,
|
8 |
+
"eval_wer": 0.40009832711037885
|
9 |
}
|
runs/Aug30_21-53-10_4b35055fdbcb/events.out.tfevents.1725061483.4b35055fdbcb.318.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bce9a3f5d305870094b0b15d0e6f2c1a88995a9bae5068167ab796c85fe77df5
|
3 |
+
size 406
|
train_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"epoch": 0.
|
3 |
-
"total_flos":
|
4 |
-
"train_loss": 0.
|
5 |
-
"train_runtime":
|
6 |
"train_samples": 124748,
|
7 |
-
"train_samples_per_second": 10.
|
8 |
-
"train_steps_per_second": 1.
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 0.6412722842118763,
|
3 |
+
"total_flos": 1.1393778193380235e+19,
|
4 |
+
"train_loss": 0.7283544036865235,
|
5 |
+
"train_runtime": 7737.7643,
|
6 |
"train_samples": 124748,
|
7 |
+
"train_samples_per_second": 10.339,
|
8 |
+
"train_steps_per_second": 1.292
|
9 |
}
|
trainer_state.json
CHANGED
@@ -1,217 +1,345 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.03206361421059382,
|
13 |
-
"grad_norm":
|
14 |
"learning_rate": 0.0002465,
|
15 |
-
"loss": 4.
|
16 |
"step": 500
|
17 |
},
|
18 |
{
|
19 |
"epoch": 0.03206361421059382,
|
20 |
-
"eval_loss": 1.
|
21 |
-
"eval_runtime":
|
22 |
-
"eval_samples_per_second":
|
23 |
-
"eval_steps_per_second":
|
24 |
-
"eval_wer": 0.
|
25 |
"step": 500
|
26 |
},
|
27 |
{
|
28 |
"epoch": 0.06412722842118763,
|
29 |
-
"grad_norm": 4.
|
30 |
-
"learning_rate": 0.
|
31 |
-
"loss": 1.
|
32 |
"step": 1000
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.06412722842118763,
|
36 |
-
"eval_loss": 1.
|
37 |
-
"eval_runtime":
|
38 |
-
"eval_samples_per_second":
|
39 |
-
"eval_steps_per_second":
|
40 |
-
"eval_wer": 0.
|
41 |
"step": 1000
|
42 |
},
|
43 |
{
|
44 |
"epoch": 0.09619084263178146,
|
45 |
-
"grad_norm": 4.
|
46 |
-
"learning_rate": 0.
|
47 |
-
"loss": 0.
|
48 |
"step": 1500
|
49 |
},
|
50 |
{
|
51 |
"epoch": 0.09619084263178146,
|
52 |
-
"eval_loss": 1.
|
53 |
-
"eval_runtime":
|
54 |
-
"eval_samples_per_second":
|
55 |
-
"eval_steps_per_second":
|
56 |
-
"eval_wer": 0.
|
57 |
"step": 1500
|
58 |
},
|
59 |
{
|
60 |
"epoch": 0.12825445684237527,
|
61 |
-
"grad_norm":
|
62 |
-
"learning_rate": 0.
|
63 |
-
"loss": 0.
|
64 |
"step": 2000
|
65 |
},
|
66 |
{
|
67 |
"epoch": 0.12825445684237527,
|
68 |
-
"eval_loss": 0.
|
69 |
-
"eval_runtime":
|
70 |
-
"eval_samples_per_second":
|
71 |
-
"eval_steps_per_second":
|
72 |
-
"eval_wer": 0.
|
73 |
"step": 2000
|
74 |
},
|
75 |
{
|
76 |
"epoch": 0.16031807105296908,
|
77 |
-
"grad_norm":
|
78 |
-
"learning_rate": 0.
|
79 |
-
"loss": 0.
|
80 |
"step": 2500
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.16031807105296908,
|
84 |
-
"eval_loss": 0.
|
85 |
-
"eval_runtime":
|
86 |
-
"eval_samples_per_second":
|
87 |
-
"eval_steps_per_second":
|
88 |
-
"eval_wer": 0.
|
89 |
"step": 2500
|
90 |
},
|
91 |
{
|
92 |
"epoch": 0.19238168526356292,
|
93 |
-
"grad_norm": 5.
|
94 |
-
"learning_rate": 0.
|
95 |
-
"loss": 0.
|
96 |
"step": 3000
|
97 |
},
|
98 |
{
|
99 |
"epoch": 0.19238168526356292,
|
100 |
-
"eval_loss": 0.
|
101 |
-
"eval_runtime":
|
102 |
-
"eval_samples_per_second":
|
103 |
-
"eval_steps_per_second":
|
104 |
-
"eval_wer": 0.
|
105 |
"step": 3000
|
106 |
},
|
107 |
{
|
108 |
"epoch": 0.22444529947415673,
|
109 |
-
"grad_norm":
|
110 |
-
"learning_rate": 0.
|
111 |
-
"loss": 0.
|
112 |
"step": 3500
|
113 |
},
|
114 |
{
|
115 |
"epoch": 0.22444529947415673,
|
116 |
-
"eval_loss": 0.
|
117 |
-
"eval_runtime":
|
118 |
-
"eval_samples_per_second":
|
119 |
-
"eval_steps_per_second":
|
120 |
-
"eval_wer": 0.
|
121 |
"step": 3500
|
122 |
},
|
123 |
{
|
124 |
"epoch": 0.25650891368475054,
|
125 |
-
"grad_norm":
|
126 |
-
"learning_rate": 0.
|
127 |
-
"loss": 0.
|
128 |
"step": 4000
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.25650891368475054,
|
132 |
-
"eval_loss": 0.
|
133 |
-
"eval_runtime":
|
134 |
-
"eval_samples_per_second":
|
135 |
-
"eval_steps_per_second":
|
136 |
-
"eval_wer": 0.
|
137 |
"step": 4000
|
138 |
},
|
139 |
{
|
140 |
"epoch": 0.2885725278953444,
|
141 |
-
"grad_norm": 9.
|
142 |
-
"learning_rate":
|
143 |
-
"loss": 0.
|
144 |
"step": 4500
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2885725278953444,
|
148 |
-
"eval_loss": 0.
|
149 |
-
"eval_runtime":
|
150 |
-
"eval_samples_per_second":
|
151 |
-
"eval_steps_per_second":
|
152 |
-
"eval_wer": 0.
|
153 |
"step": 4500
|
154 |
},
|
155 |
{
|
156 |
"epoch": 0.32063614210593816,
|
157 |
-
"grad_norm":
|
158 |
-
"learning_rate":
|
159 |
-
"loss": 0.
|
160 |
"step": 5000
|
161 |
},
|
162 |
{
|
163 |
"epoch": 0.32063614210593816,
|
164 |
-
"eval_loss": 0.
|
165 |
-
"eval_runtime":
|
166 |
-
"eval_samples_per_second":
|
167 |
-
"eval_steps_per_second":
|
168 |
-
"eval_wer": 0.
|
169 |
"step": 5000
|
170 |
},
|
171 |
{
|
172 |
"epoch": 0.352699756316532,
|
173 |
-
"grad_norm":
|
174 |
-
"learning_rate":
|
175 |
-
"loss": 0.
|
176 |
"step": 5500
|
177 |
},
|
178 |
{
|
179 |
"epoch": 0.352699756316532,
|
180 |
-
"eval_loss": 0.
|
181 |
-
"eval_runtime":
|
182 |
-
"eval_samples_per_second":
|
183 |
-
"eval_steps_per_second":
|
184 |
-
"eval_wer": 0.
|
185 |
"step": 5500
|
186 |
},
|
187 |
{
|
188 |
"epoch": 0.38476337052712584,
|
189 |
-
"grad_norm":
|
190 |
-
"learning_rate":
|
191 |
-
"loss": 0.
|
192 |
"step": 6000
|
193 |
},
|
194 |
{
|
195 |
"epoch": 0.38476337052712584,
|
196 |
-
"eval_loss": 0.
|
197 |
-
"eval_runtime":
|
198 |
-
"eval_samples_per_second":
|
199 |
-
"eval_steps_per_second":
|
200 |
-
"eval_wer": 0.
|
201 |
"step": 6000
|
202 |
},
|
203 |
{
|
204 |
-
"epoch": 0.
|
205 |
-
"
|
206 |
-
"
|
207 |
-
"
|
208 |
-
"
|
209 |
-
|
210 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
211 |
}
|
212 |
],
|
213 |
"logging_steps": 500,
|
214 |
-
"max_steps":
|
215 |
"num_input_tokens_seen": 0,
|
216 |
"num_train_epochs": 1,
|
217 |
"save_steps": 500,
|
@@ -227,7 +355,7 @@
|
|
227 |
"attributes": {}
|
228 |
}
|
229 |
},
|
230 |
-
"total_flos":
|
231 |
"train_batch_size": 8,
|
232 |
"trial_name": null,
|
233 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.6412722842118763,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 10000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.03206361421059382,
|
13 |
+
"grad_norm": 11.533838272094727,
|
14 |
"learning_rate": 0.0002465,
|
15 |
+
"loss": 4.6618,
|
16 |
"step": 500
|
17 |
},
|
18 |
{
|
19 |
"epoch": 0.03206361421059382,
|
20 |
+
"eval_loss": 1.5996026992797852,
|
21 |
+
"eval_runtime": 185.1067,
|
22 |
+
"eval_samples_per_second": 37.935,
|
23 |
+
"eval_steps_per_second": 0.594,
|
24 |
+
"eval_wer": 0.9161163448889834,
|
25 |
"step": 500
|
26 |
},
|
27 |
{
|
28 |
"epoch": 0.06412722842118763,
|
29 |
+
"grad_norm": 4.801280975341797,
|
30 |
+
"learning_rate": 0.0002874574468085106,
|
31 |
+
"loss": 1.0278,
|
32 |
"step": 1000
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.06412722842118763,
|
36 |
+
"eval_loss": 1.1463252305984497,
|
37 |
+
"eval_runtime": 184.8935,
|
38 |
+
"eval_samples_per_second": 37.979,
|
39 |
+
"eval_steps_per_second": 0.595,
|
40 |
+
"eval_wer": 0.7792157748574922,
|
41 |
"step": 1000
|
42 |
},
|
43 |
{
|
44 |
"epoch": 0.09619084263178146,
|
45 |
+
"grad_norm": 4.746730327606201,
|
46 |
+
"learning_rate": 0.0002715,
|
47 |
+
"loss": 0.8164,
|
48 |
"step": 1500
|
49 |
},
|
50 |
{
|
51 |
"epoch": 0.09619084263178146,
|
52 |
+
"eval_loss": 1.0590689182281494,
|
53 |
+
"eval_runtime": 185.3865,
|
54 |
+
"eval_samples_per_second": 37.878,
|
55 |
+
"eval_steps_per_second": 0.593,
|
56 |
+
"eval_wer": 0.7363238948165668,
|
57 |
"step": 1500
|
58 |
},
|
59 |
{
|
60 |
"epoch": 0.12825445684237527,
|
61 |
+
"grad_norm": 21.847057342529297,
|
62 |
+
"learning_rate": 0.00025554255319148935,
|
63 |
+
"loss": 0.7124,
|
64 |
"step": 2000
|
65 |
},
|
66 |
{
|
67 |
"epoch": 0.12825445684237527,
|
68 |
+
"eval_loss": 0.9373884797096252,
|
69 |
+
"eval_runtime": 186.5741,
|
70 |
+
"eval_samples_per_second": 37.637,
|
71 |
+
"eval_steps_per_second": 0.59,
|
72 |
+
"eval_wer": 0.6622596632960842,
|
73 |
"step": 2000
|
74 |
},
|
75 |
{
|
76 |
"epoch": 0.16031807105296908,
|
77 |
+
"grad_norm": 4.238761901855469,
|
78 |
+
"learning_rate": 0.0002395851063829787,
|
79 |
+
"loss": 0.6566,
|
80 |
"step": 2500
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.16031807105296908,
|
84 |
+
"eval_loss": 0.8721805810928345,
|
85 |
+
"eval_runtime": 187.0392,
|
86 |
+
"eval_samples_per_second": 37.543,
|
87 |
+
"eval_steps_per_second": 0.588,
|
88 |
+
"eval_wer": 0.615168949893036,
|
89 |
"step": 2500
|
90 |
},
|
91 |
{
|
92 |
"epoch": 0.19238168526356292,
|
93 |
+
"grad_norm": 5.65878963470459,
|
94 |
+
"learning_rate": 0.00022362765957446805,
|
95 |
+
"loss": 0.6101,
|
96 |
"step": 3000
|
97 |
},
|
98 |
{
|
99 |
"epoch": 0.19238168526356292,
|
100 |
+
"eval_loss": 0.8194323182106018,
|
101 |
+
"eval_runtime": 187.1888,
|
102 |
+
"eval_samples_per_second": 37.513,
|
103 |
+
"eval_steps_per_second": 0.588,
|
104 |
+
"eval_wer": 0.5927264513214366,
|
105 |
"step": 3000
|
106 |
},
|
107 |
{
|
108 |
"epoch": 0.22444529947415673,
|
109 |
+
"grad_norm": 6.294471740722656,
|
110 |
+
"learning_rate": 0.00020767021276595744,
|
111 |
+
"loss": 0.5777,
|
112 |
"step": 3500
|
113 |
},
|
114 |
{
|
115 |
"epoch": 0.22444529947415673,
|
116 |
+
"eval_loss": 0.7799355387687683,
|
117 |
+
"eval_runtime": 187.6349,
|
118 |
+
"eval_samples_per_second": 37.424,
|
119 |
+
"eval_steps_per_second": 0.586,
|
120 |
+
"eval_wer": 0.5707490134070343,
|
121 |
"step": 3500
|
122 |
},
|
123 |
{
|
124 |
"epoch": 0.25650891368475054,
|
125 |
+
"grad_norm": 3.848400115966797,
|
126 |
+
"learning_rate": 0.0001917446808510638,
|
127 |
+
"loss": 0.5431,
|
128 |
"step": 4000
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.25650891368475054,
|
132 |
+
"eval_loss": 0.7453157901763916,
|
133 |
+
"eval_runtime": 187.5991,
|
134 |
+
"eval_samples_per_second": 37.431,
|
135 |
+
"eval_steps_per_second": 0.586,
|
136 |
+
"eval_wer": 0.550233194700966,
|
137 |
"step": 4000
|
138 |
},
|
139 |
{
|
140 |
"epoch": 0.2885725278953444,
|
141 |
+
"grad_norm": 9.198270797729492,
|
142 |
+
"learning_rate": 0.00017585106382978722,
|
143 |
+
"loss": 0.512,
|
144 |
"step": 4500
|
145 |
},
|
146 |
{
|
147 |
"epoch": 0.2885725278953444,
|
148 |
+
"eval_loss": 0.7209311127662659,
|
149 |
+
"eval_runtime": 188.2502,
|
150 |
+
"eval_samples_per_second": 37.301,
|
151 |
+
"eval_steps_per_second": 0.584,
|
152 |
+
"eval_wer": 0.5346735938558843,
|
153 |
"step": 4500
|
154 |
},
|
155 |
{
|
156 |
"epoch": 0.32063614210593816,
|
157 |
+
"grad_norm": 7.854990482330322,
|
158 |
+
"learning_rate": 0.00015989361702127658,
|
159 |
+
"loss": 0.4953,
|
160 |
"step": 5000
|
161 |
},
|
162 |
{
|
163 |
"epoch": 0.32063614210593816,
|
164 |
+
"eval_loss": 0.6942155957221985,
|
165 |
+
"eval_runtime": 187.3133,
|
166 |
+
"eval_samples_per_second": 37.488,
|
167 |
+
"eval_steps_per_second": 0.587,
|
168 |
+
"eval_wer": 0.5225022920846676,
|
169 |
"step": 5000
|
170 |
},
|
171 |
{
|
172 |
"epoch": 0.352699756316532,
|
173 |
+
"grad_norm": 10.654298782348633,
|
174 |
+
"learning_rate": 0.00014393617021276595,
|
175 |
+
"loss": 0.4746,
|
176 |
"step": 5500
|
177 |
},
|
178 |
{
|
179 |
"epoch": 0.352699756316532,
|
180 |
+
"eval_loss": 0.6680377721786499,
|
181 |
+
"eval_runtime": 188.6546,
|
182 |
+
"eval_samples_per_second": 37.221,
|
183 |
+
"eval_steps_per_second": 0.583,
|
184 |
+
"eval_wer": 0.495714798230112,
|
185 |
"step": 5500
|
186 |
},
|
187 |
{
|
188 |
"epoch": 0.38476337052712584,
|
189 |
+
"grad_norm": 19.936861038208008,
|
190 |
+
"learning_rate": 0.00012801063829787234,
|
191 |
+
"loss": 0.4535,
|
192 |
"step": 6000
|
193 |
},
|
194 |
{
|
195 |
"epoch": 0.38476337052712584,
|
196 |
+
"eval_loss": 0.6434958577156067,
|
197 |
+
"eval_runtime": 188.0277,
|
198 |
+
"eval_samples_per_second": 37.346,
|
199 |
+
"eval_steps_per_second": 0.585,
|
200 |
+
"eval_wer": 0.47308627539563375,
|
201 |
"step": 6000
|
202 |
},
|
203 |
{
|
204 |
+
"epoch": 0.4168269847377196,
|
205 |
+
"grad_norm": 6.950649261474609,
|
206 |
+
"learning_rate": 0.0001120531914893617,
|
207 |
+
"loss": 0.4249,
|
208 |
+
"step": 6500
|
209 |
+
},
|
210 |
+
{
|
211 |
+
"epoch": 0.4168269847377196,
|
212 |
+
"eval_loss": 0.6293027400970459,
|
213 |
+
"eval_runtime": 187.8772,
|
214 |
+
"eval_samples_per_second": 37.375,
|
215 |
+
"eval_steps_per_second": 0.585,
|
216 |
+
"eval_wer": 0.4816832538301067,
|
217 |
+
"step": 6500
|
218 |
+
},
|
219 |
+
{
|
220 |
+
"epoch": 0.44889059894831346,
|
221 |
+
"grad_norm": 5.574142932891846,
|
222 |
+
"learning_rate": 9.612765957446806e-05,
|
223 |
+
"loss": 0.4065,
|
224 |
+
"step": 7000
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 0.44889059894831346,
|
228 |
+
"eval_loss": 0.5997486710548401,
|
229 |
+
"eval_runtime": 188.7633,
|
230 |
+
"eval_samples_per_second": 37.2,
|
231 |
+
"eval_steps_per_second": 0.583,
|
232 |
+
"eval_wer": 0.44947448145736724,
|
233 |
+
"step": 7000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.4809542131589073,
|
237 |
+
"grad_norm": 5.602737903594971,
|
238 |
+
"learning_rate": 8.017021276595744e-05,
|
239 |
+
"loss": 0.393,
|
240 |
+
"step": 7500
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.4809542131589073,
|
244 |
+
"eval_loss": 0.5802670121192932,
|
245 |
+
"eval_runtime": 189.0166,
|
246 |
+
"eval_samples_per_second": 37.15,
|
247 |
+
"eval_steps_per_second": 0.582,
|
248 |
+
"eval_wer": 0.44260487117819797,
|
249 |
+
"step": 7500
|
250 |
+
},
|
251 |
+
{
|
252 |
+
"epoch": 0.5130178273695011,
|
253 |
+
"grad_norm": 15.019088745117188,
|
254 |
+
"learning_rate": 6.424468085106383e-05,
|
255 |
+
"loss": 0.3808,
|
256 |
+
"step": 8000
|
257 |
+
},
|
258 |
+
{
|
259 |
+
"epoch": 0.5130178273695011,
|
260 |
+
"eval_loss": 0.5661880970001221,
|
261 |
+
"eval_runtime": 189.1322,
|
262 |
+
"eval_samples_per_second": 37.127,
|
263 |
+
"eval_steps_per_second": 0.582,
|
264 |
+
"eval_wer": 0.4299153589603901,
|
265 |
+
"step": 8000
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 0.5450814415800949,
|
269 |
+
"grad_norm": 3.79109525680542,
|
270 |
+
"learning_rate": 4.8287234042553194e-05,
|
271 |
+
"loss": 0.3722,
|
272 |
+
"step": 8500
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 0.5450814415800949,
|
276 |
+
"eval_loss": 0.553141176700592,
|
277 |
+
"eval_runtime": 188.8506,
|
278 |
+
"eval_samples_per_second": 37.183,
|
279 |
+
"eval_steps_per_second": 0.582,
|
280 |
+
"eval_wer": 0.4194182755550831,
|
281 |
+
"step": 8500
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 0.5771450557906888,
|
285 |
+
"grad_norm": 12.16059398651123,
|
286 |
+
"learning_rate": 3.232978723404255e-05,
|
287 |
+
"loss": 0.3622,
|
288 |
+
"step": 9000
|
289 |
+
},
|
290 |
+
{
|
291 |
+
"epoch": 0.5771450557906888,
|
292 |
+
"eval_loss": 0.5399113297462463,
|
293 |
+
"eval_runtime": 188.6804,
|
294 |
+
"eval_samples_per_second": 37.216,
|
295 |
+
"eval_steps_per_second": 0.583,
|
296 |
+
"eval_wer": 0.40727354867856336,
|
297 |
+
"step": 9000
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 0.6092086700012825,
|
301 |
+
"grad_norm": 7.863190650939941,
|
302 |
+
"learning_rate": 1.6372340425531912e-05,
|
303 |
+
"loss": 0.3526,
|
304 |
+
"step": 9500
|
305 |
+
},
|
306 |
+
{
|
307 |
+
"epoch": 0.6092086700012825,
|
308 |
+
"eval_loss": 0.5277913808822632,
|
309 |
+
"eval_runtime": 188.1946,
|
310 |
+
"eval_samples_per_second": 37.312,
|
311 |
+
"eval_steps_per_second": 0.585,
|
312 |
+
"eval_wer": 0.40279567892212226,
|
313 |
+
"step": 9500
|
314 |
+
},
|
315 |
+
{
|
316 |
+
"epoch": 0.6412722842118763,
|
317 |
+
"grad_norm": 5.124056339263916,
|
318 |
+
"learning_rate": 4.1489361702127654e-07,
|
319 |
+
"loss": 0.3337,
|
320 |
+
"step": 10000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 0.6412722842118763,
|
324 |
+
"eval_loss": 0.5226185917854309,
|
325 |
+
"eval_runtime": 188.4518,
|
326 |
+
"eval_samples_per_second": 37.262,
|
327 |
+
"eval_steps_per_second": 0.584,
|
328 |
+
"eval_wer": 0.40009832711037885,
|
329 |
+
"step": 10000
|
330 |
+
},
|
331 |
+
{
|
332 |
+
"epoch": 0.6412722842118763,
|
333 |
+
"step": 10000,
|
334 |
+
"total_flos": 1.1393778193380235e+19,
|
335 |
+
"train_loss": 0.7283544036865235,
|
336 |
+
"train_runtime": 7737.7643,
|
337 |
+
"train_samples_per_second": 10.339,
|
338 |
+
"train_steps_per_second": 1.292
|
339 |
}
|
340 |
],
|
341 |
"logging_steps": 500,
|
342 |
+
"max_steps": 10000,
|
343 |
"num_input_tokens_seen": 0,
|
344 |
"num_train_epochs": 1,
|
345 |
"save_steps": 500,
|
|
|
355 |
"attributes": {}
|
356 |
}
|
357 |
},
|
358 |
+
"total_flos": 1.1393778193380235e+19,
|
359 |
"train_batch_size": 8,
|
360 |
"trial_name": null,
|
361 |
"trial_params": null
|