Final model
Browse files- README.md +44 -44
- predictions.csv +0 -0
- pytorch_model.bin +1 -1
- tokenizer_config.json +1 -1
- trainer_state.json +1213 -0
README.md
CHANGED
|
@@ -26,7 +26,7 @@ model-index:
|
|
| 26 |
metrics:
|
| 27 |
- name: Test WER
|
| 28 |
type: wer
|
| 29 |
-
value:
|
| 30 |
|
| 31 |
---
|
| 32 |
|
|
@@ -136,65 +136,65 @@ for i in max_items:
|
|
| 136 |
|
| 137 |
**Output:**
|
| 138 |
```text
|
| 139 |
-
reference:
|
| 140 |
-
predicted:
|
| 141 |
---
|
| 142 |
-
reference:
|
| 143 |
-
predicted:
|
| 144 |
---
|
| 145 |
-
reference:
|
| 146 |
-
predicted:
|
| 147 |
---
|
| 148 |
-
reference:
|
| 149 |
-
predicted:
|
| 150 |
---
|
| 151 |
-
reference:
|
| 152 |
-
predicted:
|
| 153 |
---
|
| 154 |
-
reference:
|
| 155 |
-
predicted:
|
| 156 |
---
|
| 157 |
-
reference:
|
| 158 |
-
predicted:
|
| 159 |
---
|
| 160 |
-
reference:
|
| 161 |
-
predicted:
|
| 162 |
---
|
| 163 |
-
reference:
|
| 164 |
-
predicted:
|
| 165 |
---
|
| 166 |
-
reference:
|
| 167 |
-
predicted:
|
| 168 |
---
|
| 169 |
-
reference:
|
| 170 |
-
predicted:
|
| 171 |
---
|
| 172 |
-
reference:
|
| 173 |
-
predicted:
|
| 174 |
---
|
| 175 |
-
reference:
|
| 176 |
-
predicted:
|
| 177 |
---
|
| 178 |
-
reference:
|
| 179 |
-
predicted:
|
| 180 |
---
|
| 181 |
-
reference:
|
| 182 |
-
predicted:
|
| 183 |
---
|
| 184 |
-
reference:
|
| 185 |
-
predicted:
|
| 186 |
---
|
| 187 |
-
reference:
|
| 188 |
-
predicted:
|
| 189 |
---
|
| 190 |
-
reference:
|
| 191 |
-
predicted:
|
| 192 |
---
|
| 193 |
-
reference:
|
| 194 |
-
predicted:
|
| 195 |
---
|
| 196 |
-
reference:
|
| 197 |
-
predicted:
|
| 198 |
---
|
| 199 |
```
|
| 200 |
|
|
@@ -264,15 +264,15 @@ print("WER: {:.2f}".format(100 * wer.compute(predictions=result["predicted"], re
|
|
| 264 |
|
| 265 |
|
| 266 |
**Test Result**:
|
| 267 |
-
- WER:
|
| 268 |
|
| 269 |
|
| 270 |
## Training & Report
|
| 271 |
The Common Voice `train`, `validation` datasets were used for training.
|
| 272 |
|
| 273 |
-
You can see the training states [here](
|
| 274 |
|
| 275 |
-
The script used for training can be found [here](
|
| 276 |
|
| 277 |
|
| 278 |
## Questions?
|
|
|
|
| 26 |
metrics:
|
| 27 |
- name: Test WER
|
| 28 |
type: wer
|
| 29 |
+
value: 09.21
|
| 30 |
|
| 31 |
---
|
| 32 |
|
|
|
|
| 136 |
|
| 137 |
**Output:**
|
| 138 |
```text
|
| 139 |
+
reference: eða eitthvað annað dýr
|
| 140 |
+
predicted: eða eitthvað annað dýr
|
| 141 |
---
|
| 142 |
+
reference: oddgerður
|
| 143 |
+
predicted: oddgerður
|
| 144 |
---
|
| 145 |
+
reference: eiðný
|
| 146 |
+
predicted: eiðný
|
| 147 |
---
|
| 148 |
+
reference: löndum
|
| 149 |
+
predicted: löndum
|
| 150 |
---
|
| 151 |
+
reference: tileinkaði bróður sínum markið
|
| 152 |
+
predicted: tileinkaði bróður sínum markið
|
| 153 |
---
|
| 154 |
+
reference: þetta er svo mikill hégómi
|
| 155 |
+
predicted: þetta er svo mikill hégómi
|
| 156 |
---
|
| 157 |
+
reference: timarit is
|
| 158 |
+
predicted: timarit is
|
| 159 |
---
|
| 160 |
+
reference: stefna strax upp aftur
|
| 161 |
+
predicted: stefna strax upp aftur
|
| 162 |
---
|
| 163 |
+
reference: brekkuflöt
|
| 164 |
+
predicted: brekkuflöt
|
| 165 |
---
|
| 166 |
+
reference: áætlunarferð frestað vegna veðurs
|
| 167 |
+
predicted: áætluna ferð frestað vegna veðurs
|
| 168 |
---
|
| 169 |
+
reference: sagði af sér vegna kláms
|
| 170 |
+
predicted: sagði af sér vegni kláms
|
| 171 |
---
|
| 172 |
+
reference: grímúlfur
|
| 173 |
+
predicted: grímúlgur
|
| 174 |
---
|
| 175 |
+
reference: lýsti sig saklausan
|
| 176 |
+
predicted: lýsti sig saklausan
|
| 177 |
---
|
| 178 |
+
reference: belgingur is
|
| 179 |
+
predicted: belgingur is
|
| 180 |
---
|
| 181 |
+
reference: sambía
|
| 182 |
+
predicted: sambía
|
| 183 |
---
|
| 184 |
+
reference: geirastöðum
|
| 185 |
+
predicted: geirastöðum
|
| 186 |
---
|
| 187 |
+
reference: varð tvisvar fyrir eigin bíl
|
| 188 |
+
predicted: var tvisvar fyrir eigin bíl
|
| 189 |
---
|
| 190 |
+
reference: reykjavöllum
|
| 191 |
+
predicted: reykjavöllum
|
| 192 |
---
|
| 193 |
+
reference: miklir menn eru þeir þremenningar
|
| 194 |
+
predicted: miklir menn eru þeir þremenningar
|
| 195 |
---
|
| 196 |
+
reference: handverkoghonnun is
|
| 197 |
+
predicted: handverkoghonnun is
|
| 198 |
---
|
| 199 |
```
|
| 200 |
|
|
|
|
| 264 |
|
| 265 |
|
| 266 |
**Test Result**:
|
| 267 |
+
- WER: 09.21%
|
| 268 |
|
| 269 |
|
| 270 |
## Training & Report
|
| 271 |
The Common Voice `train`, `validation` datasets were used for training.
|
| 272 |
|
| 273 |
+
You can see the training states [here](https://wandb.ai/m3hrdadfi/wav2vec2_large_xlsr_is/reports/Fine-Tuning-for-Wav2Vec2-Large-XLSR-Icelandic--Vmlldzo2Mjk3ODc?accessToken=j7neoz71mce1fkzt0bch4j0l50witnmme07xe90nvs769kjjtbwneu2wfz3oip16)
|
| 274 |
|
| 275 |
+
The script used for training can be found [here](https://colab.research.google.com/github/m3hrdadfi/notebooks/blob/main/Fine_Tune_XLSR_Wav2Vec2_on_Icelandic_ASR_with_%F0%9F%A4%97_Transformers_ipynb.ipynb)
|
| 276 |
|
| 277 |
|
| 278 |
## Questions?
|
predictions.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1262097815
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31e85e7374a7ea111ba06e445df4fa572c0dfe68011d136ba65d6ec10dca3329
|
| 3 |
size 1262097815
|
tokenizer_config.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|"}
|
|
|
|
| 1 |
+
{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|", "special_tokens_map_file": "/workspace/output_models/is/xlsr-is/special_tokens_map.json", "tokenizer_file": null, "name_or_path": "/workspace/output_models/is/xlsr-is"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1213 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 25780,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.15,
|
| 12 |
+
"learning_rate": 0.00024390399999999997,
|
| 13 |
+
"loss": 10.7678,
|
| 14 |
+
"step": 300
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.15,
|
| 18 |
+
"eval_loss": 3.0923821926116943,
|
| 19 |
+
"eval_runtime": 369.7958,
|
| 20 |
+
"eval_samples_per_second": 14.681,
|
| 21 |
+
"eval_wer": 1.0,
|
| 22 |
+
"step": 300
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"epoch": 0.29,
|
| 26 |
+
"learning_rate": 0.0004113554106910039,
|
| 27 |
+
"loss": 1.4821,
|
| 28 |
+
"step": 600
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"epoch": 0.29,
|
| 32 |
+
"eval_loss": 0.4990185499191284,
|
| 33 |
+
"eval_runtime": 398.0715,
|
| 34 |
+
"eval_samples_per_second": 13.638,
|
| 35 |
+
"eval_wer": 0.6425968693867077,
|
| 36 |
+
"step": 600
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 0.44,
|
| 40 |
+
"learning_rate": 0.0004093410691003911,
|
| 41 |
+
"loss": 0.4925,
|
| 42 |
+
"step": 900
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.44,
|
| 46 |
+
"eval_loss": 0.3599244952201843,
|
| 47 |
+
"eval_runtime": 395.2724,
|
| 48 |
+
"eval_samples_per_second": 13.735,
|
| 49 |
+
"eval_wer": 0.522966384398255,
|
| 50 |
+
"step": 900
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.58,
|
| 54 |
+
"learning_rate": 0.00040732672750977834,
|
| 55 |
+
"loss": 0.3926,
|
| 56 |
+
"step": 1200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.58,
|
| 60 |
+
"eval_loss": 0.27538031339645386,
|
| 61 |
+
"eval_runtime": 370.4342,
|
| 62 |
+
"eval_samples_per_second": 14.656,
|
| 63 |
+
"eval_wer": 0.4267898383371824,
|
| 64 |
+
"step": 1200
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.73,
|
| 68 |
+
"learning_rate": 0.00040531238591916556,
|
| 69 |
+
"loss": 0.3404,
|
| 70 |
+
"step": 1500
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 0.73,
|
| 74 |
+
"eval_loss": 0.25353798270225525,
|
| 75 |
+
"eval_runtime": 396.0187,
|
| 76 |
+
"eval_samples_per_second": 13.709,
|
| 77 |
+
"eval_wer": 0.3899409802412112,
|
| 78 |
+
"step": 1500
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 0.87,
|
| 82 |
+
"learning_rate": 0.0004032980443285528,
|
| 83 |
+
"loss": 0.3148,
|
| 84 |
+
"step": 1800
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.87,
|
| 88 |
+
"eval_loss": 0.23486939072608948,
|
| 89 |
+
"eval_runtime": 394.9332,
|
| 90 |
+
"eval_samples_per_second": 13.747,
|
| 91 |
+
"eval_wer": 0.38388503977418525,
|
| 92 |
+
"step": 1800
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.02,
|
| 96 |
+
"learning_rate": 0.00040128370273794,
|
| 97 |
+
"loss": 0.2853,
|
| 98 |
+
"step": 2100
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.02,
|
| 102 |
+
"eval_loss": 0.20920990407466888,
|
| 103 |
+
"eval_runtime": 393.4003,
|
| 104 |
+
"eval_samples_per_second": 13.8,
|
| 105 |
+
"eval_wer": 0.32840646651270206,
|
| 106 |
+
"step": 2100
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 1.16,
|
| 110 |
+
"learning_rate": 0.00039926936114732727,
|
| 111 |
+
"loss": 0.2404,
|
| 112 |
+
"step": 2400
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.16,
|
| 116 |
+
"eval_loss": 0.19739878177642822,
|
| 117 |
+
"eval_runtime": 367.4384,
|
| 118 |
+
"eval_samples_per_second": 14.775,
|
| 119 |
+
"eval_wer": 0.3072106748781114,
|
| 120 |
+
"step": 2400
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 1.31,
|
| 124 |
+
"learning_rate": 0.00039725501955671443,
|
| 125 |
+
"loss": 0.2325,
|
| 126 |
+
"step": 2700
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.31,
|
| 130 |
+
"eval_loss": 0.19151932001113892,
|
| 131 |
+
"eval_runtime": 398.5555,
|
| 132 |
+
"eval_samples_per_second": 13.622,
|
| 133 |
+
"eval_wer": 0.3065434949961509,
|
| 134 |
+
"step": 2700
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.45,
|
| 138 |
+
"learning_rate": 0.00039524067796610165,
|
| 139 |
+
"loss": 0.2325,
|
| 140 |
+
"step": 3000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.45,
|
| 144 |
+
"eval_loss": 0.18169228732585907,
|
| 145 |
+
"eval_runtime": 393.6421,
|
| 146 |
+
"eval_samples_per_second": 13.792,
|
| 147 |
+
"eval_wer": 0.2896587118296125,
|
| 148 |
+
"step": 3000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 1.6,
|
| 152 |
+
"learning_rate": 0.00039322633637548887,
|
| 153 |
+
"loss": 0.2271,
|
| 154 |
+
"step": 3300
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.6,
|
| 158 |
+
"eval_loss": 0.1825951784849167,
|
| 159 |
+
"eval_runtime": 368.1876,
|
| 160 |
+
"eval_samples_per_second": 14.745,
|
| 161 |
+
"eval_wer": 0.29001796253528356,
|
| 162 |
+
"step": 3300
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.75,
|
| 166 |
+
"learning_rate": 0.00039121199478487614,
|
| 167 |
+
"loss": 0.2263,
|
| 168 |
+
"step": 3600
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.75,
|
| 172 |
+
"eval_loss": 0.1739877462387085,
|
| 173 |
+
"eval_runtime": 397.804,
|
| 174 |
+
"eval_samples_per_second": 13.647,
|
| 175 |
+
"eval_wer": 0.2704131383115217,
|
| 176 |
+
"step": 3600
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.89,
|
| 180 |
+
"learning_rate": 0.00038919765319426336,
|
| 181 |
+
"loss": 0.2187,
|
| 182 |
+
"step": 3900
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.89,
|
| 186 |
+
"eval_loss": 0.16721388697624207,
|
| 187 |
+
"eval_runtime": 396.6468,
|
| 188 |
+
"eval_samples_per_second": 13.687,
|
| 189 |
+
"eval_wer": 0.2596869386707724,
|
| 190 |
+
"step": 3900
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 2.04,
|
| 194 |
+
"learning_rate": 0.0003871833116036506,
|
| 195 |
+
"loss": 0.2015,
|
| 196 |
+
"step": 4200
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 2.04,
|
| 200 |
+
"eval_loss": 0.1631636619567871,
|
| 201 |
+
"eval_runtime": 391.6354,
|
| 202 |
+
"eval_samples_per_second": 13.862,
|
| 203 |
+
"eval_wer": 0.25517064408519374,
|
| 204 |
+
"step": 4200
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 2.18,
|
| 208 |
+
"learning_rate": 0.0003851689700130378,
|
| 209 |
+
"loss": 0.1792,
|
| 210 |
+
"step": 4500
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 2.18,
|
| 214 |
+
"eval_loss": 0.17116306722164154,
|
| 215 |
+
"eval_runtime": 368.9065,
|
| 216 |
+
"eval_samples_per_second": 14.716,
|
| 217 |
+
"eval_wer": 0.26646138054914037,
|
| 218 |
+
"step": 4500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.33,
|
| 222 |
+
"learning_rate": 0.000383154628422425,
|
| 223 |
+
"loss": 0.1773,
|
| 224 |
+
"step": 4800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.33,
|
| 228 |
+
"eval_loss": 0.16152945160865784,
|
| 229 |
+
"eval_runtime": 398.4133,
|
| 230 |
+
"eval_samples_per_second": 13.627,
|
| 231 |
+
"eval_wer": 0.24146779574031307,
|
| 232 |
+
"step": 4800
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 2.47,
|
| 236 |
+
"learning_rate": 0.00038114028683181223,
|
| 237 |
+
"loss": 0.179,
|
| 238 |
+
"step": 5100
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 2.47,
|
| 242 |
+
"eval_loss": 0.15521417558193207,
|
| 243 |
+
"eval_runtime": 396.2358,
|
| 244 |
+
"eval_samples_per_second": 13.701,
|
| 245 |
+
"eval_wer": 0.24613805491403642,
|
| 246 |
+
"step": 5100
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 2.62,
|
| 250 |
+
"learning_rate": 0.00037912594524119945,
|
| 251 |
+
"loss": 0.1717,
|
| 252 |
+
"step": 5400
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 2.62,
|
| 256 |
+
"eval_loss": 0.15384772419929504,
|
| 257 |
+
"eval_runtime": 393.2934,
|
| 258 |
+
"eval_samples_per_second": 13.804,
|
| 259 |
+
"eval_wer": 0.2374133949191686,
|
| 260 |
+
"step": 5400
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.76,
|
| 264 |
+
"learning_rate": 0.0003771116036505867,
|
| 265 |
+
"loss": 0.1744,
|
| 266 |
+
"step": 5700
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.76,
|
| 270 |
+
"eval_loss": 0.14799486100673676,
|
| 271 |
+
"eval_runtime": 370.8447,
|
| 272 |
+
"eval_samples_per_second": 14.64,
|
| 273 |
+
"eval_wer": 0.23828586091865536,
|
| 274 |
+
"step": 5700
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 2.91,
|
| 278 |
+
"learning_rate": 0.0003750972620599739,
|
| 279 |
+
"loss": 0.166,
|
| 280 |
+
"step": 6000
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 2.91,
|
| 284 |
+
"eval_loss": 0.1445951759815216,
|
| 285 |
+
"eval_runtime": 393.1105,
|
| 286 |
+
"eval_samples_per_second": 13.81,
|
| 287 |
+
"eval_wer": 0.22992045162945857,
|
| 288 |
+
"step": 6000
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 3.06,
|
| 292 |
+
"learning_rate": 0.0003730829204693611,
|
| 293 |
+
"loss": 0.1532,
|
| 294 |
+
"step": 6300
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 3.06,
|
| 298 |
+
"eval_loss": 0.1482468694448471,
|
| 299 |
+
"eval_runtime": 395.8029,
|
| 300 |
+
"eval_samples_per_second": 13.716,
|
| 301 |
+
"eval_wer": 0.22037464716448552,
|
| 302 |
+
"step": 6300
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 3.2,
|
| 306 |
+
"learning_rate": 0.00037106857887874833,
|
| 307 |
+
"loss": 0.142,
|
| 308 |
+
"step": 6600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 3.2,
|
| 312 |
+
"eval_loss": 0.15104271471500397,
|
| 313 |
+
"eval_runtime": 369.3329,
|
| 314 |
+
"eval_samples_per_second": 14.699,
|
| 315 |
+
"eval_wer": 0.228175519630485,
|
| 316 |
+
"step": 6600
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 3.35,
|
| 320 |
+
"learning_rate": 0.0003690542372881356,
|
| 321 |
+
"loss": 0.1406,
|
| 322 |
+
"step": 6900
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 3.35,
|
| 326 |
+
"eval_loss": 0.1345045417547226,
|
| 327 |
+
"eval_runtime": 401.0837,
|
| 328 |
+
"eval_samples_per_second": 13.536,
|
| 329 |
+
"eval_wer": 0.20795483705414422,
|
| 330 |
+
"step": 6900
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 3.49,
|
| 334 |
+
"learning_rate": 0.0003670398956975228,
|
| 335 |
+
"loss": 0.1458,
|
| 336 |
+
"step": 7200
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.49,
|
| 340 |
+
"eval_loss": 0.13937360048294067,
|
| 341 |
+
"eval_runtime": 390.779,
|
| 342 |
+
"eval_samples_per_second": 13.893,
|
| 343 |
+
"eval_wer": 0.2111367718758019,
|
| 344 |
+
"step": 7200
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 3.64,
|
| 348 |
+
"learning_rate": 0.00036502555410691004,
|
| 349 |
+
"loss": 0.1449,
|
| 350 |
+
"step": 7500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 3.64,
|
| 354 |
+
"eval_loss": 0.14066852629184723,
|
| 355 |
+
"eval_runtime": 395.9112,
|
| 356 |
+
"eval_samples_per_second": 13.713,
|
| 357 |
+
"eval_wer": 0.20846805234795998,
|
| 358 |
+
"step": 7500
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 3.78,
|
| 362 |
+
"learning_rate": 0.00036301121251629726,
|
| 363 |
+
"loss": 0.137,
|
| 364 |
+
"step": 7800
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 3.78,
|
| 368 |
+
"eval_loss": 0.1394476294517517,
|
| 369 |
+
"eval_runtime": 367.4977,
|
| 370 |
+
"eval_samples_per_second": 14.773,
|
| 371 |
+
"eval_wer": 0.20744162176032846,
|
| 372 |
+
"step": 7800
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 3.93,
|
| 376 |
+
"learning_rate": 0.0003609968709256845,
|
| 377 |
+
"loss": 0.1388,
|
| 378 |
+
"step": 8100
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 3.93,
|
| 382 |
+
"eval_loss": 0.1351165473461151,
|
| 383 |
+
"eval_runtime": 397.2254,
|
| 384 |
+
"eval_samples_per_second": 13.667,
|
| 385 |
+
"eval_wer": 0.19964074929432898,
|
| 386 |
+
"step": 8100
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 4.07,
|
| 390 |
+
"learning_rate": 0.0003589825293350717,
|
| 391 |
+
"loss": 0.1307,
|
| 392 |
+
"step": 8400
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 4.07,
|
| 396 |
+
"eval_loss": 0.1401081681251526,
|
| 397 |
+
"eval_runtime": 396.6542,
|
| 398 |
+
"eval_samples_per_second": 13.687,
|
| 399 |
+
"eval_wer": 0.20446497305619707,
|
| 400 |
+
"step": 8400
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 4.22,
|
| 404 |
+
"learning_rate": 0.0003569681877444589,
|
| 405 |
+
"loss": 0.122,
|
| 406 |
+
"step": 8700
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 4.22,
|
| 410 |
+
"eval_loss": 0.14438064396381378,
|
| 411 |
+
"eval_runtime": 389.1513,
|
| 412 |
+
"eval_samples_per_second": 13.951,
|
| 413 |
+
"eval_wer": 0.20369515011547343,
|
| 414 |
+
"step": 8700
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 4.36,
|
| 418 |
+
"learning_rate": 0.0003549538461538462,
|
| 419 |
+
"loss": 0.1193,
|
| 420 |
+
"step": 9000
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 4.36,
|
| 424 |
+
"eval_loss": 0.14555728435516357,
|
| 425 |
+
"eval_runtime": 367.57,
|
| 426 |
+
"eval_samples_per_second": 14.77,
|
| 427 |
+
"eval_wer": 0.21524249422632794,
|
| 428 |
+
"step": 9000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 4.51,
|
| 432 |
+
"learning_rate": 0.00035293950456323335,
|
| 433 |
+
"loss": 0.1344,
|
| 434 |
+
"step": 9300
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 4.51,
|
| 438 |
+
"eval_loss": 0.1305903196334839,
|
| 439 |
+
"eval_runtime": 395.5909,
|
| 440 |
+
"eval_samples_per_second": 13.724,
|
| 441 |
+
"eval_wer": 0.19522709776751349,
|
| 442 |
+
"step": 9300
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 4.66,
|
| 446 |
+
"learning_rate": 0.00035092516297262057,
|
| 447 |
+
"loss": 0.1267,
|
| 448 |
+
"step": 9600
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 4.66,
|
| 452 |
+
"eval_loss": 0.1303185522556305,
|
| 453 |
+
"eval_runtime": 392.9371,
|
| 454 |
+
"eval_samples_per_second": 13.816,
|
| 455 |
+
"eval_wer": 0.18963305106492173,
|
| 456 |
+
"step": 9600
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 4.8,
|
| 460 |
+
"learning_rate": 0.0003489108213820078,
|
| 461 |
+
"loss": 0.1214,
|
| 462 |
+
"step": 9900
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 4.8,
|
| 466 |
+
"eval_loss": 0.1294514238834381,
|
| 467 |
+
"eval_runtime": 369.1477,
|
| 468 |
+
"eval_samples_per_second": 14.707,
|
| 469 |
+
"eval_wer": 0.19240441365152683,
|
| 470 |
+
"step": 9900
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.95,
|
| 474 |
+
"learning_rate": 0.00034689647979139506,
|
| 475 |
+
"loss": 0.1179,
|
| 476 |
+
"step": 10200
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 4.95,
|
| 480 |
+
"eval_loss": 0.12323788553476334,
|
| 481 |
+
"eval_runtime": 391.972,
|
| 482 |
+
"eval_samples_per_second": 13.85,
|
| 483 |
+
"eval_wer": 0.19183987682832948,
|
| 484 |
+
"step": 10200
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 5.09,
|
| 488 |
+
"learning_rate": 0.0003448821382007823,
|
| 489 |
+
"loss": 0.0998,
|
| 490 |
+
"step": 10500
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 5.09,
|
| 494 |
+
"eval_loss": 0.12908631563186646,
|
| 495 |
+
"eval_runtime": 393.594,
|
| 496 |
+
"eval_samples_per_second": 13.793,
|
| 497 |
+
"eval_wer": 0.18516807800872467,
|
| 498 |
+
"step": 10500
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 5.24,
|
| 502 |
+
"learning_rate": 0.0003428677966101695,
|
| 503 |
+
"loss": 0.1017,
|
| 504 |
+
"step": 10800
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 5.24,
|
| 508 |
+
"eval_loss": 0.1282510757446289,
|
| 509 |
+
"eval_runtime": 393.5495,
|
| 510 |
+
"eval_samples_per_second": 13.795,
|
| 511 |
+
"eval_wer": 0.1841929689504747,
|
| 512 |
+
"step": 10800
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 5.38,
|
| 516 |
+
"learning_rate": 0.00034085345501955666,
|
| 517 |
+
"loss": 0.1024,
|
| 518 |
+
"step": 11100
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 5.38,
|
| 522 |
+
"eval_loss": 0.12217988073825836,
|
| 523 |
+
"eval_runtime": 368.9647,
|
| 524 |
+
"eval_samples_per_second": 14.714,
|
| 525 |
+
"eval_wer": 0.1739286630741596,
|
| 526 |
+
"step": 11100
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 5.53,
|
| 530 |
+
"learning_rate": 0.00033883911342894393,
|
| 531 |
+
"loss": 0.1042,
|
| 532 |
+
"step": 11400
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 5.53,
|
| 536 |
+
"eval_loss": 0.1310078352689743,
|
| 537 |
+
"eval_runtime": 394.8352,
|
| 538 |
+
"eval_samples_per_second": 13.75,
|
| 539 |
+
"eval_wer": 0.18137028483448805,
|
| 540 |
+
"step": 11400
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 5.67,
|
| 544 |
+
"learning_rate": 0.00033682477183833115,
|
| 545 |
+
"loss": 0.1096,
|
| 546 |
+
"step": 11700
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 5.67,
|
| 550 |
+
"eval_loss": 0.12405824661254883,
|
| 551 |
+
"eval_runtime": 392.553,
|
| 552 |
+
"eval_samples_per_second": 13.83,
|
| 553 |
+
"eval_wer": 0.17890685142417245,
|
| 554 |
+
"step": 11700
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 5.82,
|
| 558 |
+
"learning_rate": 0.0003348171447196871,
|
| 559 |
+
"loss": 0.1091,
|
| 560 |
+
"step": 12000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 5.82,
|
| 564 |
+
"eval_loss": 0.12054669111967087,
|
| 565 |
+
"eval_runtime": 396.4385,
|
| 566 |
+
"eval_samples_per_second": 13.694,
|
| 567 |
+
"eval_wer": 0.17942006671798819,
|
| 568 |
+
"step": 12000
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 5.96,
|
| 572 |
+
"learning_rate": 0.0003328028031290743,
|
| 573 |
+
"loss": 0.1039,
|
| 574 |
+
"step": 12300
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 5.96,
|
| 578 |
+
"eval_loss": 0.11941063404083252,
|
| 579 |
+
"eval_runtime": 368.032,
|
| 580 |
+
"eval_samples_per_second": 14.751,
|
| 581 |
+
"eval_wer": 0.17829099307159355,
|
| 582 |
+
"step": 12300
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 6.11,
|
| 586 |
+
"learning_rate": 0.0003307884615384615,
|
| 587 |
+
"loss": 0.0963,
|
| 588 |
+
"step": 12600
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 6.11,
|
| 592 |
+
"eval_loss": 0.13149844110012054,
|
| 593 |
+
"eval_runtime": 395.5282,
|
| 594 |
+
"eval_samples_per_second": 13.726,
|
| 595 |
+
"eval_wer": 0.17823967154221196,
|
| 596 |
+
"step": 12600
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 6.26,
|
| 600 |
+
"learning_rate": 0.00032877411994784874,
|
| 601 |
+
"loss": 0.0947,
|
| 602 |
+
"step": 12900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 6.26,
|
| 606 |
+
"eval_loss": 0.11894866079092026,
|
| 607 |
+
"eval_runtime": 394.9332,
|
| 608 |
+
"eval_samples_per_second": 13.747,
|
| 609 |
+
"eval_wer": 0.16925840390043623,
|
| 610 |
+
"step": 12900
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 6.4,
|
| 614 |
+
"learning_rate": 0.00032675977835723596,
|
| 615 |
+
"loss": 0.0946,
|
| 616 |
+
"step": 13200
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 6.4,
|
| 620 |
+
"eval_loss": 0.12667880952358246,
|
| 621 |
+
"eval_runtime": 365.722,
|
| 622 |
+
"eval_samples_per_second": 14.845,
|
| 623 |
+
"eval_wer": 0.1762894534257121,
|
| 624 |
+
"step": 13200
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 6.55,
|
| 628 |
+
"learning_rate": 0.0003247454367666232,
|
| 629 |
+
"loss": 0.0935,
|
| 630 |
+
"step": 13500
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 6.55,
|
| 634 |
+
"eval_loss": 0.12555569410324097,
|
| 635 |
+
"eval_runtime": 401.4704,
|
| 636 |
+
"eval_samples_per_second": 13.523,
|
| 637 |
+
"eval_wer": 0.1729022324865281,
|
| 638 |
+
"step": 13500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 6.69,
|
| 642 |
+
"learning_rate": 0.00032273780964797916,
|
| 643 |
+
"loss": 0.088,
|
| 644 |
+
"step": 13800
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 6.69,
|
| 648 |
+
"eval_loss": 0.12047957628965378,
|
| 649 |
+
"eval_runtime": 389.5023,
|
| 650 |
+
"eval_samples_per_second": 13.938,
|
| 651 |
+
"eval_wer": 0.16546061072619964,
|
| 652 |
+
"step": 13800
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 6.84,
|
| 656 |
+
"learning_rate": 0.00032072346805736633,
|
| 657 |
+
"loss": 0.0947,
|
| 658 |
+
"step": 14100
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 6.84,
|
| 662 |
+
"eval_loss": 0.1205919086933136,
|
| 663 |
+
"eval_runtime": 397.9241,
|
| 664 |
+
"eval_samples_per_second": 13.643,
|
| 665 |
+
"eval_wer": 0.17202976648704132,
|
| 666 |
+
"step": 14100
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 6.98,
|
| 670 |
+
"learning_rate": 0.00031870912646675355,
|
| 671 |
+
"loss": 0.0897,
|
| 672 |
+
"step": 14400
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 6.98,
|
| 676 |
+
"eval_loss": 0.1126471757888794,
|
| 677 |
+
"eval_runtime": 364.5953,
|
| 678 |
+
"eval_samples_per_second": 14.89,
|
| 679 |
+
"eval_wer": 0.16556325378496278,
|
| 680 |
+
"step": 14400
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 7.13,
|
| 684 |
+
"learning_rate": 0.00031669478487614077,
|
| 685 |
+
"loss": 0.0853,
|
| 686 |
+
"step": 14700
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 7.13,
|
| 690 |
+
"eval_loss": 0.1229565218091011,
|
| 691 |
+
"eval_runtime": 386.8026,
|
| 692 |
+
"eval_samples_per_second": 14.036,
|
| 693 |
+
"eval_wer": 0.1657685399024891,
|
| 694 |
+
"step": 14700
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 7.27,
|
| 698 |
+
"learning_rate": 0.00031468044328552804,
|
| 699 |
+
"loss": 0.0844,
|
| 700 |
+
"step": 15000
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 7.27,
|
| 704 |
+
"eval_loss": 0.1236405223608017,
|
| 705 |
+
"eval_runtime": 399.2408,
|
| 706 |
+
"eval_samples_per_second": 13.598,
|
| 707 |
+
"eval_wer": 0.15991788555298947,
|
| 708 |
+
"step": 15000
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 7.42,
|
| 712 |
+
"learning_rate": 0.00031266610169491526,
|
| 713 |
+
"loss": 0.0836,
|
| 714 |
+
"step": 15300
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 7.42,
|
| 718 |
+
"eval_loss": 0.12277714908123016,
|
| 719 |
+
"eval_runtime": 395.0189,
|
| 720 |
+
"eval_samples_per_second": 13.744,
|
| 721 |
+
"eval_wer": 0.17233769566333076,
|
| 722 |
+
"step": 15300
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 6.05,
|
| 726 |
+
"learning_rate": 0.00016603860759493668,
|
| 727 |
+
"loss": 0.0827,
|
| 728 |
+
"step": 15600
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 6.05,
|
| 732 |
+
"eval_loss": 0.11456963419914246,
|
| 733 |
+
"eval_runtime": 391.5783,
|
| 734 |
+
"eval_samples_per_second": 13.864,
|
| 735 |
+
"eval_wer": 0.14801129073646394,
|
| 736 |
+
"step": 15600
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 6.17,
|
| 740 |
+
"learning_rate": 0.00016114936708860758,
|
| 741 |
+
"loss": 0.0774,
|
| 742 |
+
"step": 15900
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 6.17,
|
| 746 |
+
"eval_loss": 0.10517676174640656,
|
| 747 |
+
"eval_runtime": 387.996,
|
| 748 |
+
"eval_samples_per_second": 13.992,
|
| 749 |
+
"eval_wer": 0.14251988709263536,
|
| 750 |
+
"step": 15900
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 6.28,
|
| 754 |
+
"learning_rate": 0.00015626012658227848,
|
| 755 |
+
"loss": 0.0662,
|
| 756 |
+
"step": 16200
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 6.28,
|
| 760 |
+
"eval_loss": 0.10210742056369781,
|
| 761 |
+
"eval_runtime": 393.9886,
|
| 762 |
+
"eval_samples_per_second": 13.78,
|
| 763 |
+
"eval_wer": 0.13456505003849115,
|
| 764 |
+
"step": 16200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 6.4,
|
| 768 |
+
"learning_rate": 0.00015138718354430379,
|
| 769 |
+
"loss": 0.0709,
|
| 770 |
+
"step": 16500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 6.4,
|
| 774 |
+
"eval_loss": 0.09889358282089233,
|
| 775 |
+
"eval_runtime": 373.1692,
|
| 776 |
+
"eval_samples_per_second": 14.548,
|
| 777 |
+
"eval_wer": 0.13451372850910956,
|
| 778 |
+
"step": 16500
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 6.52,
|
| 782 |
+
"learning_rate": 0.00014649794303797466,
|
| 783 |
+
"loss": 0.068,
|
| 784 |
+
"step": 16800
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 6.52,
|
| 788 |
+
"eval_loss": 0.09926428645849228,
|
| 789 |
+
"eval_runtime": 387.0555,
|
| 790 |
+
"eval_samples_per_second": 14.026,
|
| 791 |
+
"eval_wer": 0.1367718758018989,
|
| 792 |
+
"step": 16800
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 6.63,
|
| 796 |
+
"learning_rate": 0.00014160870253164559,
|
| 797 |
+
"loss": 0.0641,
|
| 798 |
+
"step": 17100
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 6.63,
|
| 802 |
+
"eval_loss": 0.09896103292703629,
|
| 803 |
+
"eval_runtime": 394.68,
|
| 804 |
+
"eval_samples_per_second": 13.755,
|
| 805 |
+
"eval_wer": 0.12948421862971515,
|
| 806 |
+
"step": 17100
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 6.75,
|
| 810 |
+
"learning_rate": 0.00013671946202531646,
|
| 811 |
+
"loss": 0.0633,
|
| 812 |
+
"step": 17400
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 6.75,
|
| 816 |
+
"eval_loss": 0.09418534487485886,
|
| 817 |
+
"eval_runtime": 389.6282,
|
| 818 |
+
"eval_samples_per_second": 13.934,
|
| 819 |
+
"eval_wer": 0.12799589427764949,
|
| 820 |
+
"step": 17400
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 6.87,
|
| 824 |
+
"learning_rate": 0.00013183022151898733,
|
| 825 |
+
"loss": 0.0662,
|
| 826 |
+
"step": 17700
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 6.87,
|
| 830 |
+
"eval_loss": 0.09439662098884583,
|
| 831 |
+
"eval_runtime": 385.8863,
|
| 832 |
+
"eval_samples_per_second": 14.069,
|
| 833 |
+
"eval_wer": 0.12553246086733386,
|
| 834 |
+
"step": 17700
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 6.98,
|
| 838 |
+
"learning_rate": 0.00012694098101265823,
|
| 839 |
+
"loss": 0.0626,
|
| 840 |
+
"step": 18000
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 6.98,
|
| 844 |
+
"eval_loss": 0.09515263140201569,
|
| 845 |
+
"eval_runtime": 386.6219,
|
| 846 |
+
"eval_samples_per_second": 14.042,
|
| 847 |
+
"eval_wer": 0.12199127534000513,
|
| 848 |
+
"step": 18000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 7.1,
|
| 852 |
+
"learning_rate": 0.00012205174050632911,
|
| 853 |
+
"loss": 0.0554,
|
| 854 |
+
"step": 18300
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 7.1,
|
| 858 |
+
"eval_loss": 0.09619747847318649,
|
| 859 |
+
"eval_runtime": 388.8787,
|
| 860 |
+
"eval_samples_per_second": 13.961,
|
| 861 |
+
"eval_wer": 0.12158070310495253,
|
| 862 |
+
"step": 18300
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 7.21,
|
| 866 |
+
"learning_rate": 0.0001171625,
|
| 867 |
+
"loss": 0.0532,
|
| 868 |
+
"step": 18600
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 7.21,
|
| 872 |
+
"eval_loss": 0.10246041417121887,
|
| 873 |
+
"eval_runtime": 392.3678,
|
| 874 |
+
"eval_samples_per_second": 13.837,
|
| 875 |
+
"eval_wer": 0.12178598922247882,
|
| 876 |
+
"step": 18600
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 7.33,
|
| 880 |
+
"learning_rate": 0.00011228955696202532,
|
| 881 |
+
"loss": 0.0522,
|
| 882 |
+
"step": 18900
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 7.33,
|
| 886 |
+
"eval_loss": 0.09738663583993912,
|
| 887 |
+
"eval_runtime": 387.5929,
|
| 888 |
+
"eval_samples_per_second": 14.007,
|
| 889 |
+
"eval_wer": 0.1191172696946369,
|
| 890 |
+
"step": 18900
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 7.45,
|
| 894 |
+
"learning_rate": 0.0001074003164556962,
|
| 895 |
+
"loss": 0.0514,
|
| 896 |
+
"step": 19200
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 7.45,
|
| 900 |
+
"eval_loss": 0.096110999584198,
|
| 901 |
+
"eval_runtime": 385.6979,
|
| 902 |
+
"eval_samples_per_second": 14.076,
|
| 903 |
+
"eval_wer": 0.12045162945855786,
|
| 904 |
+
"step": 19200
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 7.56,
|
| 908 |
+
"learning_rate": 0.00010251107594936708,
|
| 909 |
+
"loss": 0.0507,
|
| 910 |
+
"step": 19500
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 7.56,
|
| 914 |
+
"eval_loss": 0.09606499969959259,
|
| 915 |
+
"eval_runtime": 388.8775,
|
| 916 |
+
"eval_samples_per_second": 13.961,
|
| 917 |
+
"eval_wer": 0.11778290993071594,
|
| 918 |
+
"step": 19500
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 7.68,
|
| 922 |
+
"learning_rate": 9.76381329113924e-05,
|
| 923 |
+
"loss": 0.0512,
|
| 924 |
+
"step": 19800
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 7.68,
|
| 928 |
+
"eval_loss": 0.0907408595085144,
|
| 929 |
+
"eval_runtime": 369.8847,
|
| 930 |
+
"eval_samples_per_second": 14.678,
|
| 931 |
+
"eval_wer": 0.11311265075699256,
|
| 932 |
+
"step": 19800
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 7.8,
|
| 936 |
+
"learning_rate": 9.274889240506329e-05,
|
| 937 |
+
"loss": 0.0488,
|
| 938 |
+
"step": 20100
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 7.8,
|
| 942 |
+
"eval_loss": 0.08931271731853485,
|
| 943 |
+
"eval_runtime": 392.0795,
|
| 944 |
+
"eval_samples_per_second": 13.847,
|
| 945 |
+
"eval_wer": 0.1107005388760585,
|
| 946 |
+
"step": 20100
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 7.91,
|
| 950 |
+
"learning_rate": 8.785965189873417e-05,
|
| 951 |
+
"loss": 0.0489,
|
| 952 |
+
"step": 20400
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 7.91,
|
| 956 |
+
"eval_loss": 0.08885154128074646,
|
| 957 |
+
"eval_runtime": 387.663,
|
| 958 |
+
"eval_samples_per_second": 14.004,
|
| 959 |
+
"eval_wer": 0.1098793944059533,
|
| 960 |
+
"step": 20400
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 8.03,
|
| 964 |
+
"learning_rate": 8.297041139240506e-05,
|
| 965 |
+
"loss": 0.0458,
|
| 966 |
+
"step": 20700
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 8.03,
|
| 970 |
+
"eval_loss": 0.0913764014840126,
|
| 971 |
+
"eval_runtime": 388.0349,
|
| 972 |
+
"eval_samples_per_second": 13.991,
|
| 973 |
+
"eval_wer": 0.10644085193738774,
|
| 974 |
+
"step": 20700
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 8.15,
|
| 978 |
+
"learning_rate": 7.808117088607595e-05,
|
| 979 |
+
"loss": 0.0432,
|
| 980 |
+
"step": 21000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 8.15,
|
| 984 |
+
"eval_loss": 0.0885995477437973,
|
| 985 |
+
"eval_runtime": 390.3347,
|
| 986 |
+
"eval_samples_per_second": 13.909,
|
| 987 |
+
"eval_wer": 0.10746728252501925,
|
| 988 |
+
"step": 21000
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 8.26,
|
| 992 |
+
"learning_rate": 7.319193037974683e-05,
|
| 993 |
+
"loss": 0.0402,
|
| 994 |
+
"step": 21300
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 8.26,
|
| 998 |
+
"eval_loss": 0.09285327792167664,
|
| 999 |
+
"eval_runtime": 390.0106,
|
| 1000 |
+
"eval_samples_per_second": 13.92,
|
| 1001 |
+
"eval_wer": 0.1060816012317167,
|
| 1002 |
+
"step": 21300
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 8.38,
|
| 1006 |
+
"learning_rate": 6.830268987341772e-05,
|
| 1007 |
+
"loss": 0.043,
|
| 1008 |
+
"step": 21600
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 8.38,
|
| 1012 |
+
"eval_loss": 0.09266022592782974,
|
| 1013 |
+
"eval_runtime": 390.9718,
|
| 1014 |
+
"eval_samples_per_second": 13.886,
|
| 1015 |
+
"eval_wer": 0.10515781370284835,
|
| 1016 |
+
"step": 21600
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 8.49,
|
| 1020 |
+
"learning_rate": 6.34134493670886e-05,
|
| 1021 |
+
"loss": 0.041,
|
| 1022 |
+
"step": 21900
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 8.49,
|
| 1026 |
+
"eval_loss": 0.0903313010931015,
|
| 1027 |
+
"eval_runtime": 392.2575,
|
| 1028 |
+
"eval_samples_per_second": 13.84,
|
| 1029 |
+
"eval_wer": 0.10325891711573004,
|
| 1030 |
+
"step": 21900
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 8.61,
|
| 1034 |
+
"learning_rate": 5.8524208860759495e-05,
|
| 1035 |
+
"loss": 0.0387,
|
| 1036 |
+
"step": 22200
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 8.61,
|
| 1040 |
+
"eval_loss": 0.09009859710931778,
|
| 1041 |
+
"eval_runtime": 381.4418,
|
| 1042 |
+
"eval_samples_per_second": 14.233,
|
| 1043 |
+
"eval_wer": 0.10212984346933539,
|
| 1044 |
+
"step": 22200
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 8.73,
|
| 1048 |
+
"learning_rate": 5.365126582278481e-05,
|
| 1049 |
+
"loss": 0.0381,
|
| 1050 |
+
"step": 22500
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 8.73,
|
| 1054 |
+
"eval_loss": 0.0909804031252861,
|
| 1055 |
+
"eval_runtime": 387.8645,
|
| 1056 |
+
"eval_samples_per_second": 13.997,
|
| 1057 |
+
"eval_wer": 0.10028226841159867,
|
| 1058 |
+
"step": 22500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 8.84,
|
| 1062 |
+
"learning_rate": 4.87620253164557e-05,
|
| 1063 |
+
"loss": 0.0368,
|
| 1064 |
+
"step": 22800
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 8.84,
|
| 1068 |
+
"eval_loss": 0.08640129864215851,
|
| 1069 |
+
"eval_runtime": 389.071,
|
| 1070 |
+
"eval_samples_per_second": 13.954,
|
| 1071 |
+
"eval_wer": 0.09899923017705928,
|
| 1072 |
+
"step": 22800
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 8.96,
|
| 1076 |
+
"learning_rate": 4.387278481012658e-05,
|
| 1077 |
+
"loss": 0.0393,
|
| 1078 |
+
"step": 23100
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 8.96,
|
| 1082 |
+
"eval_loss": 0.08331574499607086,
|
| 1083 |
+
"eval_runtime": 370.5127,
|
| 1084 |
+
"eval_samples_per_second": 14.653,
|
| 1085 |
+
"eval_wer": 0.0987426225301514,
|
| 1086 |
+
"step": 23100
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 9.08,
|
| 1090 |
+
"learning_rate": 3.898354430379747e-05,
|
| 1091 |
+
"loss": 0.0384,
|
| 1092 |
+
"step": 23400
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 9.08,
|
| 1096 |
+
"eval_loss": 0.08789246529340744,
|
| 1097 |
+
"eval_runtime": 384.7677,
|
| 1098 |
+
"eval_samples_per_second": 14.11,
|
| 1099 |
+
"eval_wer": 0.09905055170644085,
|
| 1100 |
+
"step": 23400
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 9.19,
|
| 1104 |
+
"learning_rate": 3.409430379746835e-05,
|
| 1105 |
+
"loss": 0.0335,
|
| 1106 |
+
"step": 23700
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 9.19,
|
| 1110 |
+
"eval_loss": 0.0857023224234581,
|
| 1111 |
+
"eval_runtime": 380.952,
|
| 1112 |
+
"eval_samples_per_second": 14.251,
|
| 1113 |
+
"eval_wer": 0.09612522453169105,
|
| 1114 |
+
"step": 23700
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 9.31,
|
| 1118 |
+
"learning_rate": 2.9205063291139242e-05,
|
| 1119 |
+
"loss": 0.0356,
|
| 1120 |
+
"step": 24000
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 9.31,
|
| 1124 |
+
"eval_loss": 0.08656106889247894,
|
| 1125 |
+
"eval_runtime": 388.804,
|
| 1126 |
+
"eval_samples_per_second": 13.963,
|
| 1127 |
+
"eval_wer": 0.0953554015909674,
|
| 1128 |
+
"step": 24000
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 9.43,
|
| 1132 |
+
"learning_rate": 2.4315822784810126e-05,
|
| 1133 |
+
"loss": 0.0354,
|
| 1134 |
+
"step": 24300
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 9.43,
|
| 1138 |
+
"eval_loss": 0.08513284474611282,
|
| 1139 |
+
"eval_runtime": 388.1691,
|
| 1140 |
+
"eval_samples_per_second": 13.986,
|
| 1141 |
+
"eval_wer": 0.09597125994354631,
|
| 1142 |
+
"step": 24300
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 9.54,
|
| 1146 |
+
"learning_rate": 1.9426582278481013e-05,
|
| 1147 |
+
"loss": 0.0311,
|
| 1148 |
+
"step": 24600
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 9.54,
|
| 1152 |
+
"eval_loss": 0.08604401350021362,
|
| 1153 |
+
"eval_runtime": 387.2711,
|
| 1154 |
+
"eval_samples_per_second": 14.019,
|
| 1155 |
+
"eval_wer": 0.09402104182704644,
|
| 1156 |
+
"step": 24600
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 9.66,
|
| 1160 |
+
"learning_rate": 1.4537341772151898e-05,
|
| 1161 |
+
"loss": 0.0328,
|
| 1162 |
+
"step": 24900
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 9.66,
|
| 1166 |
+
"eval_loss": 0.08348560333251953,
|
| 1167 |
+
"eval_runtime": 387.9108,
|
| 1168 |
+
"eval_samples_per_second": 13.995,
|
| 1169 |
+
"eval_wer": 0.09412368488580959,
|
| 1170 |
+
"step": 24900
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 9.78,
|
| 1174 |
+
"learning_rate": 9.648101265822784e-06,
|
| 1175 |
+
"loss": 0.0284,
|
| 1176 |
+
"step": 25200
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 9.78,
|
| 1180 |
+
"eval_loss": 0.0844394788146019,
|
| 1181 |
+
"eval_runtime": 389.4719,
|
| 1182 |
+
"eval_samples_per_second": 13.939,
|
| 1183 |
+
"eval_wer": 0.09319989735694123,
|
| 1184 |
+
"step": 25200
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 9.89,
|
| 1188 |
+
"learning_rate": 4.758860759493671e-06,
|
| 1189 |
+
"loss": 0.0305,
|
| 1190 |
+
"step": 25500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 9.89,
|
| 1194 |
+
"eval_loss": 0.08396408706903458,
|
| 1195 |
+
"eval_runtime": 389.8127,
|
| 1196 |
+
"eval_samples_per_second": 13.927,
|
| 1197 |
+
"eval_wer": 0.09196818065178343,
|
| 1198 |
+
"step": 25500
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 10.0,
|
| 1202 |
+
"step": 25780,
|
| 1203 |
+
"total_flos": 1.6778924993531563e+20,
|
| 1204 |
+
"train_runtime": 66940.1986,
|
| 1205 |
+
"train_samples_per_second": 0.385
|
| 1206 |
+
}
|
| 1207 |
+
],
|
| 1208 |
+
"max_steps": 25780,
|
| 1209 |
+
"num_train_epochs": 10,
|
| 1210 |
+
"total_flos": 1.6778924993531563e+20,
|
| 1211 |
+
"trial_name": null,
|
| 1212 |
+
"trial_params": null
|
| 1213 |
+
}
|