End of training
Browse files- README.md +1 -1
- all_results.json +27 -0
- eval_djelia_bm-text-normalization-benchmark__level-1__test_results.json +9 -0
- eval_djelia_bm-text-normalization__bamadaba__dev__500__results.json +9 -0
- eval_oza75_bm-text-normalization__cbr__dev__500__results.json +9 -0
- runs/Jan13_01-20-54_a5f52b8b320c/events.out.tfevents.1768272752.a5f52b8b320c.4991.1 +3 -0
- train_results.json +9 -0
- trainer_state.json +1151 -0
README.md
CHANGED
|
@@ -14,7 +14,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
|
| 15 |
# bm-byt5-text-normalization-9
|
| 16 |
|
| 17 |
-
This model is a fine-tuned version of [oza75/byt5-bambara-pretrained](https://huggingface.co/oza75/byt5-bambara-pretrained) on
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
|
|
| 14 |
|
| 15 |
# bm-byt5-text-normalization-9
|
| 16 |
|
| 17 |
+
This model is a fine-tuned version of [oza75/byt5-bambara-pretrained](https://huggingface.co/oza75/byt5-bambara-pretrained) on the oza75/bm-text-normalization and the djelia/bm-text-normalization datasets.
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
all_results.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.07097326367155066,
|
| 4 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.08283345401287079,
|
| 5 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 234.0095,
|
| 6 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.137,
|
| 7 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.269,
|
| 8 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.16193917494730503,
|
| 9 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.07097326367155066,
|
| 10 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.08283345401287079,
|
| 11 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 227.408,
|
| 12 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 2.199,
|
| 13 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 0.277,
|
| 14 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.16193917494730503,
|
| 15 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.07097326367155066,
|
| 16 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.08283345401287079,
|
| 17 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 228.3101,
|
| 18 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.19,
|
| 19 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.276,
|
| 20 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.16193917494730503,
|
| 21 |
+
"total_flos": 1.1016496820713882e+17,
|
| 22 |
+
"train_loss": 0.1502396887776639,
|
| 23 |
+
"train_runtime": 5207.863,
|
| 24 |
+
"train_samples": 99883,
|
| 25 |
+
"train_samples_per_second": 19.179,
|
| 26 |
+
"train_steps_per_second": 2.398
|
| 27 |
+
}
|
eval_djelia_bm-text-normalization-benchmark__level-1__test_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.07097326367155066,
|
| 4 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.08283345401287079,
|
| 5 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 234.0095,
|
| 6 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.137,
|
| 7 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.269,
|
| 8 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.16193917494730503
|
| 9 |
+
}
|
eval_djelia_bm-text-normalization__bamadaba__dev__500__results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.07097326367155066,
|
| 4 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.08283345401287079,
|
| 5 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 227.408,
|
| 6 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 2.199,
|
| 7 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 0.277,
|
| 8 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.16193917494730503
|
| 9 |
+
}
|
eval_oza75_bm-text-normalization__cbr__dev__500__results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.07097326367155066,
|
| 4 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.08283345401287079,
|
| 5 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 228.3101,
|
| 6 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.19,
|
| 7 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.276,
|
| 8 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.16193917494730503
|
| 9 |
+
}
|
runs/Jan13_01-20-54_a5f52b8b320c/events.out.tfevents.1768272752.a5f52b8b320c.4991.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c9399f058817b43e69a286e29fac6c9559311ff60e75b71ffb4d5b5c8ea5e75
|
| 3 |
+
size 2071
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 1.1016496820713882e+17,
|
| 4 |
+
"train_loss": 0.1502396887776639,
|
| 5 |
+
"train_runtime": 5207.863,
|
| 6 |
+
"train_samples": 99883,
|
| 7 |
+
"train_samples_per_second": 19.179,
|
| 8 |
+
"train_steps_per_second": 2.398
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 12000,
|
| 3 |
+
"best_metric": 0.07097326367155066,
|
| 4 |
+
"best_model_checkpoint": "./output/bm-byt5-text-normalization-9/checkpoint-12000",
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 1500,
|
| 7 |
+
"global_step": 12486,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.008008970046452027,
|
| 14 |
+
"grad_norm": 3.8596339225769043,
|
| 15 |
+
"learning_rate": 7.926341072858287e-06,
|
| 16 |
+
"loss": 1.568,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.016017940092904054,
|
| 21 |
+
"grad_norm": 1.0137945413589478,
|
| 22 |
+
"learning_rate": 1.593274619695757e-05,
|
| 23 |
+
"loss": 0.8709,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.024026910139356077,
|
| 28 |
+
"grad_norm": 0.7859073877334595,
|
| 29 |
+
"learning_rate": 2.3939151321056846e-05,
|
| 30 |
+
"loss": 0.6016,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.03203588018580811,
|
| 35 |
+
"grad_norm": 0.6444743871688843,
|
| 36 |
+
"learning_rate": 3.194555644515613e-05,
|
| 37 |
+
"loss": 0.5168,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.040044850232260135,
|
| 42 |
+
"grad_norm": 0.4445948302745819,
|
| 43 |
+
"learning_rate": 3.995196156925541e-05,
|
| 44 |
+
"loss": 0.4461,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.048053820278712155,
|
| 49 |
+
"grad_norm": 0.41510170698165894,
|
| 50 |
+
"learning_rate": 4.795836669335468e-05,
|
| 51 |
+
"loss": 0.3826,
|
| 52 |
+
"step": 600
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.05606279032516418,
|
| 56 |
+
"grad_norm": 0.3964841365814209,
|
| 57 |
+
"learning_rate": 5.5964771817453964e-05,
|
| 58 |
+
"loss": 0.3543,
|
| 59 |
+
"step": 700
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.06407176037161622,
|
| 63 |
+
"grad_norm": 0.3848731815814972,
|
| 64 |
+
"learning_rate": 6.397117694155325e-05,
|
| 65 |
+
"loss": 0.3161,
|
| 66 |
+
"step": 800
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.07208073041806824,
|
| 70 |
+
"grad_norm": 0.3667680025100708,
|
| 71 |
+
"learning_rate": 7.197758206565253e-05,
|
| 72 |
+
"loss": 0.2937,
|
| 73 |
+
"step": 900
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.08008970046452027,
|
| 77 |
+
"grad_norm": 0.4185062050819397,
|
| 78 |
+
"learning_rate": 7.99839871897518e-05,
|
| 79 |
+
"loss": 0.2748,
|
| 80 |
+
"step": 1000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.08809867051097228,
|
| 84 |
+
"grad_norm": 0.565210223197937,
|
| 85 |
+
"learning_rate": 8.799039231385109e-05,
|
| 86 |
+
"loss": 0.2425,
|
| 87 |
+
"step": 1100
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.09610764055742431,
|
| 91 |
+
"grad_norm": 0.3510247468948364,
|
| 92 |
+
"learning_rate": 9.599679743795037e-05,
|
| 93 |
+
"loss": 0.2429,
|
| 94 |
+
"step": 1200
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.10411661060387634,
|
| 98 |
+
"grad_norm": 0.33350569009780884,
|
| 99 |
+
"learning_rate": 9.955504138115156e-05,
|
| 100 |
+
"loss": 0.2319,
|
| 101 |
+
"step": 1300
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.11212558065032836,
|
| 105 |
+
"grad_norm": 0.2997024357318878,
|
| 106 |
+
"learning_rate": 9.866512414345466e-05,
|
| 107 |
+
"loss": 0.219,
|
| 108 |
+
"step": 1400
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.12013455069678039,
|
| 112 |
+
"grad_norm": 0.336683988571167,
|
| 113 |
+
"learning_rate": 9.777520690575776e-05,
|
| 114 |
+
"loss": 0.2154,
|
| 115 |
+
"step": 1500
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.12013455069678039,
|
| 119 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.4623481471072902,
|
| 120 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.18778733909130096,
|
| 121 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 233.0229,
|
| 122 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.146,
|
| 123 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.27,
|
| 124 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.6122854561878952,
|
| 125 |
+
"step": 1500
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.12013455069678039,
|
| 129 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.5852987060658897,
|
| 130 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.6063626408576965,
|
| 131 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 69.2146,
|
| 132 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 7.051,
|
| 133 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 0.881,
|
| 134 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.8124754805806198,
|
| 135 |
+
"step": 1500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.12013455069678039,
|
| 139 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5726033834586466,
|
| 140 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5617944598197937,
|
| 141 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 28.2591,
|
| 142 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.866,
|
| 143 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.389,
|
| 144 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.8147345612134345,
|
| 145 |
+
"step": 1500
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.12814352074323243,
|
| 149 |
+
"grad_norm": 0.3779921233654022,
|
| 150 |
+
"learning_rate": 9.688528966806088e-05,
|
| 151 |
+
"loss": 0.1948,
|
| 152 |
+
"step": 1600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.13615249078968444,
|
| 156 |
+
"grad_norm": 0.32918840646743774,
|
| 157 |
+
"learning_rate": 9.599537243036398e-05,
|
| 158 |
+
"loss": 0.1989,
|
| 159 |
+
"step": 1700
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.14416146083613648,
|
| 163 |
+
"grad_norm": 0.26091116666793823,
|
| 164 |
+
"learning_rate": 9.510545519266708e-05,
|
| 165 |
+
"loss": 0.185,
|
| 166 |
+
"step": 1800
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 0.1521704308825885,
|
| 170 |
+
"grad_norm": 0.3381774425506592,
|
| 171 |
+
"learning_rate": 9.42155379549702e-05,
|
| 172 |
+
"loss": 0.1835,
|
| 173 |
+
"step": 1900
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.16017940092904054,
|
| 177 |
+
"grad_norm": 0.4004645347595215,
|
| 178 |
+
"learning_rate": 9.33256207172733e-05,
|
| 179 |
+
"loss": 0.1685,
|
| 180 |
+
"step": 2000
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 0.16818837097549255,
|
| 184 |
+
"grad_norm": 0.22172071039676666,
|
| 185 |
+
"learning_rate": 9.243570347957641e-05,
|
| 186 |
+
"loss": 0.1749,
|
| 187 |
+
"step": 2100
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.17619734102194456,
|
| 191 |
+
"grad_norm": 0.35416314005851746,
|
| 192 |
+
"learning_rate": 9.154578624187951e-05,
|
| 193 |
+
"loss": 0.1642,
|
| 194 |
+
"step": 2200
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.1842063110683966,
|
| 198 |
+
"grad_norm": 0.2757333219051361,
|
| 199 |
+
"learning_rate": 9.065586900418261e-05,
|
| 200 |
+
"loss": 0.1714,
|
| 201 |
+
"step": 2300
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.19221528111484862,
|
| 205 |
+
"grad_norm": 0.27378392219543457,
|
| 206 |
+
"learning_rate": 8.976595176648572e-05,
|
| 207 |
+
"loss": 0.1663,
|
| 208 |
+
"step": 2400
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.20022425116130066,
|
| 212 |
+
"grad_norm": 0.34761300683021545,
|
| 213 |
+
"learning_rate": 8.887603452878883e-05,
|
| 214 |
+
"loss": 0.1544,
|
| 215 |
+
"step": 2500
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.20823322120775267,
|
| 219 |
+
"grad_norm": 0.3532973825931549,
|
| 220 |
+
"learning_rate": 8.798611729109193e-05,
|
| 221 |
+
"loss": 0.1522,
|
| 222 |
+
"step": 2600
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.2162421912542047,
|
| 226 |
+
"grad_norm": 0.3141867220401764,
|
| 227 |
+
"learning_rate": 8.709620005339504e-05,
|
| 228 |
+
"loss": 0.1594,
|
| 229 |
+
"step": 2700
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.22425116130065673,
|
| 233 |
+
"grad_norm": 0.3187774419784546,
|
| 234 |
+
"learning_rate": 8.620628281569814e-05,
|
| 235 |
+
"loss": 0.1534,
|
| 236 |
+
"step": 2800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.23226013134710877,
|
| 240 |
+
"grad_norm": 0.32024797797203064,
|
| 241 |
+
"learning_rate": 8.531636557800124e-05,
|
| 242 |
+
"loss": 0.157,
|
| 243 |
+
"step": 2900
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.24026910139356078,
|
| 247 |
+
"grad_norm": 0.3410184979438782,
|
| 248 |
+
"learning_rate": 8.442644834030436e-05,
|
| 249 |
+
"loss": 0.1497,
|
| 250 |
+
"step": 3000
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.24026910139356078,
|
| 254 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.21550017245123462,
|
| 255 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.13786430656909943,
|
| 256 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 237.885,
|
| 257 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.102,
|
| 258 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.265,
|
| 259 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.33604336043360433,
|
| 260 |
+
"step": 3000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.24026910139356078,
|
| 264 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.296411856474259,
|
| 265 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.5071771740913391,
|
| 266 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 48.4541,
|
| 267 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 10.071,
|
| 268 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.259,
|
| 269 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.5237347979599843,
|
| 270 |
+
"step": 3000
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.24026910139356078,
|
| 274 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.7525845864661654,
|
| 275 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5753647089004517,
|
| 276 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 32.5425,
|
| 277 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.489,
|
| 278 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.338,
|
| 279 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.6338028169014085,
|
| 280 |
+
"step": 3000
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 0.24827807144001282,
|
| 284 |
+
"grad_norm": 0.2752779722213745,
|
| 285 |
+
"learning_rate": 8.353653110260746e-05,
|
| 286 |
+
"loss": 0.141,
|
| 287 |
+
"step": 3100
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 0.25628704148646486,
|
| 291 |
+
"grad_norm": 0.33353230357170105,
|
| 292 |
+
"learning_rate": 8.264661386491057e-05,
|
| 293 |
+
"loss": 0.1505,
|
| 294 |
+
"step": 3200
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.2642960115329169,
|
| 298 |
+
"grad_norm": 0.21438618004322052,
|
| 299 |
+
"learning_rate": 8.175669662721367e-05,
|
| 300 |
+
"loss": 0.1399,
|
| 301 |
+
"step": 3300
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 0.2723049815793689,
|
| 305 |
+
"grad_norm": 0.38063517212867737,
|
| 306 |
+
"learning_rate": 8.086677938951677e-05,
|
| 307 |
+
"loss": 0.1471,
|
| 308 |
+
"step": 3400
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.2803139516258209,
|
| 312 |
+
"grad_norm": 0.27488982677459717,
|
| 313 |
+
"learning_rate": 7.997686215181989e-05,
|
| 314 |
+
"loss": 0.1319,
|
| 315 |
+
"step": 3500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.28832292167227297,
|
| 319 |
+
"grad_norm": 0.33845141530036926,
|
| 320 |
+
"learning_rate": 7.9086944914123e-05,
|
| 321 |
+
"loss": 0.1322,
|
| 322 |
+
"step": 3600
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 0.296331891718725,
|
| 326 |
+
"grad_norm": 0.8517484068870544,
|
| 327 |
+
"learning_rate": 7.819702767642609e-05,
|
| 328 |
+
"loss": 0.1353,
|
| 329 |
+
"step": 3700
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.304340861765177,
|
| 333 |
+
"grad_norm": 0.20209847390651703,
|
| 334 |
+
"learning_rate": 7.73071104387292e-05,
|
| 335 |
+
"loss": 0.124,
|
| 336 |
+
"step": 3800
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 0.312349831811629,
|
| 340 |
+
"grad_norm": 0.42707929015159607,
|
| 341 |
+
"learning_rate": 7.64171932010323e-05,
|
| 342 |
+
"loss": 0.1266,
|
| 343 |
+
"step": 3900
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"epoch": 0.3203588018580811,
|
| 347 |
+
"grad_norm": 0.34921887516975403,
|
| 348 |
+
"learning_rate": 7.55272759633354e-05,
|
| 349 |
+
"loss": 0.1226,
|
| 350 |
+
"step": 4000
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.3283677719045331,
|
| 354 |
+
"grad_norm": 0.33792367577552795,
|
| 355 |
+
"learning_rate": 7.463735872563852e-05,
|
| 356 |
+
"loss": 0.1188,
|
| 357 |
+
"step": 4100
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.3363767419509851,
|
| 361 |
+
"grad_norm": 0.28311803936958313,
|
| 362 |
+
"learning_rate": 7.374744148794162e-05,
|
| 363 |
+
"loss": 0.1167,
|
| 364 |
+
"step": 4200
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"epoch": 0.3443857119974371,
|
| 368 |
+
"grad_norm": 0.24853067100048065,
|
| 369 |
+
"learning_rate": 7.285752425024473e-05,
|
| 370 |
+
"loss": 0.1193,
|
| 371 |
+
"step": 4300
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.35239468204388913,
|
| 375 |
+
"grad_norm": 0.2954888641834259,
|
| 376 |
+
"learning_rate": 7.196760701254784e-05,
|
| 377 |
+
"loss": 0.1164,
|
| 378 |
+
"step": 4400
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 0.3604036520903412,
|
| 382 |
+
"grad_norm": 0.3384742736816406,
|
| 383 |
+
"learning_rate": 7.107768977485094e-05,
|
| 384 |
+
"loss": 0.1107,
|
| 385 |
+
"step": 4500
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.3604036520903412,
|
| 389 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.13066693918219793,
|
| 390 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.10385449975728989,
|
| 391 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 239.7727,
|
| 392 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.085,
|
| 393 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.263,
|
| 394 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.23932550436615477,
|
| 395 |
+
"step": 4500
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.3604036520903412,
|
| 399 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.11562815453794623,
|
| 400 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.3626447021961212,
|
| 401 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 40.1376,
|
| 402 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 12.158,
|
| 403 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.52,
|
| 404 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.29619458611220084,
|
| 405 |
+
"step": 4500
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.3604036520903412,
|
| 409 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5462875939849624,
|
| 410 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5127649903297424,
|
| 411 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 28.2264,
|
| 412 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.87,
|
| 413 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.39,
|
| 414 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.47670639219934996,
|
| 415 |
+
"step": 4500
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.3684126221367932,
|
| 419 |
+
"grad_norm": 0.23822121322155,
|
| 420 |
+
"learning_rate": 7.018777253715405e-05,
|
| 421 |
+
"loss": 0.1111,
|
| 422 |
+
"step": 4600
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3764215921832452,
|
| 426 |
+
"grad_norm": 0.26078513264656067,
|
| 427 |
+
"learning_rate": 6.929785529945715e-05,
|
| 428 |
+
"loss": 0.1107,
|
| 429 |
+
"step": 4700
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.38443056222969724,
|
| 433 |
+
"grad_norm": 0.23475381731987,
|
| 434 |
+
"learning_rate": 6.840793806176025e-05,
|
| 435 |
+
"loss": 0.1103,
|
| 436 |
+
"step": 4800
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.3924395322761493,
|
| 440 |
+
"grad_norm": 0.2875431478023529,
|
| 441 |
+
"learning_rate": 6.751802082406337e-05,
|
| 442 |
+
"loss": 0.1125,
|
| 443 |
+
"step": 4900
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4004485023226013,
|
| 447 |
+
"grad_norm": 0.21920998394489288,
|
| 448 |
+
"learning_rate": 6.662810358636648e-05,
|
| 449 |
+
"loss": 0.1051,
|
| 450 |
+
"step": 5000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.40845747236905333,
|
| 454 |
+
"grad_norm": 0.3680271804332733,
|
| 455 |
+
"learning_rate": 6.573818634866957e-05,
|
| 456 |
+
"loss": 0.1125,
|
| 457 |
+
"step": 5100
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.41646644241550534,
|
| 461 |
+
"grad_norm": 0.2292424738407135,
|
| 462 |
+
"learning_rate": 6.484826911097268e-05,
|
| 463 |
+
"loss": 0.0995,
|
| 464 |
+
"step": 5200
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4244754124619574,
|
| 468 |
+
"grad_norm": 0.23008547723293304,
|
| 469 |
+
"learning_rate": 6.39583518732758e-05,
|
| 470 |
+
"loss": 0.1026,
|
| 471 |
+
"step": 5300
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4324843825084094,
|
| 475 |
+
"grad_norm": 0.34461820125579834,
|
| 476 |
+
"learning_rate": 6.30684346355789e-05,
|
| 477 |
+
"loss": 0.1062,
|
| 478 |
+
"step": 5400
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.44049335255486144,
|
| 482 |
+
"grad_norm": 0.287975937128067,
|
| 483 |
+
"learning_rate": 6.2178517397882e-05,
|
| 484 |
+
"loss": 0.1009,
|
| 485 |
+
"step": 5500
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.44850232260131345,
|
| 489 |
+
"grad_norm": 0.31931379437446594,
|
| 490 |
+
"learning_rate": 6.12886001601851e-05,
|
| 491 |
+
"loss": 0.1021,
|
| 492 |
+
"step": 5600
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.4565112926477655,
|
| 496 |
+
"grad_norm": 0.2849687933921814,
|
| 497 |
+
"learning_rate": 6.039868292248821e-05,
|
| 498 |
+
"loss": 0.1075,
|
| 499 |
+
"step": 5700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.46452026269421753,
|
| 503 |
+
"grad_norm": 0.22519993782043457,
|
| 504 |
+
"learning_rate": 5.9508765684791314e-05,
|
| 505 |
+
"loss": 0.1115,
|
| 506 |
+
"step": 5800
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.47252923274066955,
|
| 510 |
+
"grad_norm": 0.16592390835285187,
|
| 511 |
+
"learning_rate": 5.861884844709442e-05,
|
| 512 |
+
"loss": 0.0957,
|
| 513 |
+
"step": 5900
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.48053820278712156,
|
| 517 |
+
"grad_norm": 0.20374052226543427,
|
| 518 |
+
"learning_rate": 5.772893120939753e-05,
|
| 519 |
+
"loss": 0.1073,
|
| 520 |
+
"step": 6000
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.48053820278712156,
|
| 524 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.11223381832581787,
|
| 525 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.09635704010725021,
|
| 526 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 237.7988,
|
| 527 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.103,
|
| 528 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.265,
|
| 529 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.21547726588376995,
|
| 530 |
+
"step": 6000
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.48053820278712156,
|
| 534 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.10736900064237864,
|
| 535 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.3359106481075287,
|
| 536 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 40.6894,
|
| 537 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 11.993,
|
| 538 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.499,
|
| 539 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.28167908983915263,
|
| 540 |
+
"step": 6000
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 0.48053820278712156,
|
| 544 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5159774436090225,
|
| 545 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.506378173828125,
|
| 546 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 28.9504,
|
| 547 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.798,
|
| 548 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.38,
|
| 549 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.42903575297941493,
|
| 550 |
+
"step": 6000
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.48854717283357363,
|
| 554 |
+
"grad_norm": 0.2751060128211975,
|
| 555 |
+
"learning_rate": 5.683901397170064e-05,
|
| 556 |
+
"loss": 0.1023,
|
| 557 |
+
"step": 6100
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.49655614288002564,
|
| 561 |
+
"grad_norm": 0.21174342930316925,
|
| 562 |
+
"learning_rate": 5.594909673400374e-05,
|
| 563 |
+
"loss": 0.1056,
|
| 564 |
+
"step": 6200
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.5045651129264777,
|
| 568 |
+
"grad_norm": 0.13474728167057037,
|
| 569 |
+
"learning_rate": 5.5059179496306845e-05,
|
| 570 |
+
"loss": 0.0935,
|
| 571 |
+
"step": 6300
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.5125740829729297,
|
| 575 |
+
"grad_norm": 0.3412100076675415,
|
| 576 |
+
"learning_rate": 5.416926225860995e-05,
|
| 577 |
+
"loss": 0.102,
|
| 578 |
+
"step": 6400
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.5205830530193817,
|
| 582 |
+
"grad_norm": 0.1809050291776657,
|
| 583 |
+
"learning_rate": 5.327934502091306e-05,
|
| 584 |
+
"loss": 0.0931,
|
| 585 |
+
"step": 6500
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.5285920230658337,
|
| 589 |
+
"grad_norm": 0.31216275691986084,
|
| 590 |
+
"learning_rate": 5.238942778321616e-05,
|
| 591 |
+
"loss": 0.0994,
|
| 592 |
+
"step": 6600
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.5366009931122857,
|
| 596 |
+
"grad_norm": 0.2171262800693512,
|
| 597 |
+
"learning_rate": 5.149951054551927e-05,
|
| 598 |
+
"loss": 0.0985,
|
| 599 |
+
"step": 6700
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.5446099631587378,
|
| 603 |
+
"grad_norm": 0.1729833483695984,
|
| 604 |
+
"learning_rate": 5.0609593307822376e-05,
|
| 605 |
+
"loss": 0.1006,
|
| 606 |
+
"step": 6800
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.5526189332051898,
|
| 610 |
+
"grad_norm": 0.3678678870201111,
|
| 611 |
+
"learning_rate": 4.9719676070125484e-05,
|
| 612 |
+
"loss": 0.0916,
|
| 613 |
+
"step": 6900
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.5606279032516418,
|
| 617 |
+
"grad_norm": 0.21691511571407318,
|
| 618 |
+
"learning_rate": 4.8829758832428584e-05,
|
| 619 |
+
"loss": 0.0909,
|
| 620 |
+
"step": 7000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.5686368732980939,
|
| 624 |
+
"grad_norm": 0.11386135965585709,
|
| 625 |
+
"learning_rate": 4.793984159473169e-05,
|
| 626 |
+
"loss": 0.1012,
|
| 627 |
+
"step": 7100
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.5766458433445459,
|
| 631 |
+
"grad_norm": 0.33011430501937866,
|
| 632 |
+
"learning_rate": 4.70499243570348e-05,
|
| 633 |
+
"loss": 0.0995,
|
| 634 |
+
"step": 7200
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.5846548133909979,
|
| 638 |
+
"grad_norm": 0.436538428068161,
|
| 639 |
+
"learning_rate": 4.616000711933791e-05,
|
| 640 |
+
"loss": 0.0934,
|
| 641 |
+
"step": 7300
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.59266378343745,
|
| 645 |
+
"grad_norm": 0.1885899156332016,
|
| 646 |
+
"learning_rate": 4.527008988164101e-05,
|
| 647 |
+
"loss": 0.0986,
|
| 648 |
+
"step": 7400
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 0.6006727534839019,
|
| 652 |
+
"grad_norm": 0.24165351688861847,
|
| 653 |
+
"learning_rate": 4.4380172643944115e-05,
|
| 654 |
+
"loss": 0.0925,
|
| 655 |
+
"step": 7500
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 0.6006727534839019,
|
| 659 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.09868042870099511,
|
| 660 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.0899689719080925,
|
| 661 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 237.6964,
|
| 662 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.104,
|
| 663 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.265,
|
| 664 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.19337548931044865,
|
| 665 |
+
"step": 7500
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.6006727534839019,
|
| 669 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.10773607414884831,
|
| 670 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.31528761982917786,
|
| 671 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 39.2229,
|
| 672 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 12.442,
|
| 673 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.555,
|
| 674 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.2793252255786583,
|
| 675 |
+
"step": 7500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.6006727534839019,
|
| 679 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5488721804511278,
|
| 680 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5185889005661011,
|
| 681 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 28.6015,
|
| 682 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.832,
|
| 683 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.385,
|
| 684 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.4528710725893825,
|
| 685 |
+
"step": 7500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 0.608681723530354,
|
| 689 |
+
"grad_norm": 0.4197905957698822,
|
| 690 |
+
"learning_rate": 4.349025540624722e-05,
|
| 691 |
+
"loss": 0.0941,
|
| 692 |
+
"step": 7600
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.6166906935768061,
|
| 696 |
+
"grad_norm": 0.16225901246070862,
|
| 697 |
+
"learning_rate": 4.2600338168550324e-05,
|
| 698 |
+
"loss": 0.0921,
|
| 699 |
+
"step": 7700
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.624699663623258,
|
| 703 |
+
"grad_norm": 0.2778748869895935,
|
| 704 |
+
"learning_rate": 4.171042093085343e-05,
|
| 705 |
+
"loss": 0.0868,
|
| 706 |
+
"step": 7800
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.6327086336697101,
|
| 710 |
+
"grad_norm": 0.25382688641548157,
|
| 711 |
+
"learning_rate": 4.082050369315654e-05,
|
| 712 |
+
"loss": 0.0934,
|
| 713 |
+
"step": 7900
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 0.6407176037161622,
|
| 717 |
+
"grad_norm": 0.3628831207752228,
|
| 718 |
+
"learning_rate": 3.9930586455459647e-05,
|
| 719 |
+
"loss": 0.0968,
|
| 720 |
+
"step": 8000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 0.6487265737626141,
|
| 724 |
+
"grad_norm": 0.24050642549991608,
|
| 725 |
+
"learning_rate": 3.904066921776275e-05,
|
| 726 |
+
"loss": 0.0911,
|
| 727 |
+
"step": 8100
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 0.6567355438090662,
|
| 731 |
+
"grad_norm": 0.22333455085754395,
|
| 732 |
+
"learning_rate": 3.8150751980065855e-05,
|
| 733 |
+
"loss": 0.0942,
|
| 734 |
+
"step": 8200
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.6647445138555181,
|
| 738 |
+
"grad_norm": 0.36757221817970276,
|
| 739 |
+
"learning_rate": 3.726083474236896e-05,
|
| 740 |
+
"loss": 0.0935,
|
| 741 |
+
"step": 8300
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.6727534839019702,
|
| 745 |
+
"grad_norm": 0.4592411518096924,
|
| 746 |
+
"learning_rate": 3.637091750467207e-05,
|
| 747 |
+
"loss": 0.0903,
|
| 748 |
+
"step": 8400
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 0.6807624539484223,
|
| 752 |
+
"grad_norm": 0.2041786164045334,
|
| 753 |
+
"learning_rate": 3.548100026697517e-05,
|
| 754 |
+
"loss": 0.0984,
|
| 755 |
+
"step": 8500
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 0.6887714239948742,
|
| 759 |
+
"grad_norm": 0.2337992787361145,
|
| 760 |
+
"learning_rate": 3.459108302927828e-05,
|
| 761 |
+
"loss": 0.0938,
|
| 762 |
+
"step": 8600
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.6967803940413263,
|
| 766 |
+
"grad_norm": 0.34608423709869385,
|
| 767 |
+
"learning_rate": 3.3701165791581386e-05,
|
| 768 |
+
"loss": 0.09,
|
| 769 |
+
"step": 8700
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 0.7047893640877783,
|
| 773 |
+
"grad_norm": 0.24163080751895905,
|
| 774 |
+
"learning_rate": 3.281124855388449e-05,
|
| 775 |
+
"loss": 0.0866,
|
| 776 |
+
"step": 8800
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.7127983341342303,
|
| 780 |
+
"grad_norm": 0.23344071209430695,
|
| 781 |
+
"learning_rate": 3.1921331316187594e-05,
|
| 782 |
+
"loss": 0.0876,
|
| 783 |
+
"step": 8900
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.7208073041806824,
|
| 787 |
+
"grad_norm": 0.33907049894332886,
|
| 788 |
+
"learning_rate": 3.10314140784907e-05,
|
| 789 |
+
"loss": 0.0904,
|
| 790 |
+
"step": 9000
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 0.7208073041806824,
|
| 794 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.08734974387798117,
|
| 795 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.0864705815911293,
|
| 796 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 239.9022,
|
| 797 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.084,
|
| 798 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.263,
|
| 799 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.18066847335140018,
|
| 800 |
+
"step": 9000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.7208073041806824,
|
| 804 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.10360649720106452,
|
| 805 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.30589863657951355,
|
| 806 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 40.4318,
|
| 807 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 12.07,
|
| 808 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.509,
|
| 809 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.2734405649274225,
|
| 810 |
+
"step": 9000
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.7208073041806824,
|
| 814 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5204417293233082,
|
| 815 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5159913301467896,
|
| 816 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 28.5651,
|
| 817 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.836,
|
| 818 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.385,
|
| 819 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.4431202600216685,
|
| 820 |
+
"step": 9000
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 0.7288162742271344,
|
| 824 |
+
"grad_norm": 0.3358537554740906,
|
| 825 |
+
"learning_rate": 3.014149684079381e-05,
|
| 826 |
+
"loss": 0.0975,
|
| 827 |
+
"step": 9100
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 0.7368252442735864,
|
| 831 |
+
"grad_norm": 0.30242183804512024,
|
| 832 |
+
"learning_rate": 2.9251579603096914e-05,
|
| 833 |
+
"loss": 0.0932,
|
| 834 |
+
"step": 9200
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.7448342143200385,
|
| 838 |
+
"grad_norm": 0.2576250731945038,
|
| 839 |
+
"learning_rate": 2.836166236540002e-05,
|
| 840 |
+
"loss": 0.0884,
|
| 841 |
+
"step": 9300
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.7528431843664904,
|
| 845 |
+
"grad_norm": 0.1822267770767212,
|
| 846 |
+
"learning_rate": 2.7471745127703125e-05,
|
| 847 |
+
"loss": 0.0887,
|
| 848 |
+
"step": 9400
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.7608521544129425,
|
| 852 |
+
"grad_norm": 0.28615352511405945,
|
| 853 |
+
"learning_rate": 2.6581827890006233e-05,
|
| 854 |
+
"loss": 0.096,
|
| 855 |
+
"step": 9500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.7688611244593945,
|
| 859 |
+
"grad_norm": 0.25158432126045227,
|
| 860 |
+
"learning_rate": 2.5691910652309337e-05,
|
| 861 |
+
"loss": 0.0885,
|
| 862 |
+
"step": 9600
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.7768700945058465,
|
| 866 |
+
"grad_norm": 0.3026082515716553,
|
| 867 |
+
"learning_rate": 2.480199341461244e-05,
|
| 868 |
+
"loss": 0.0877,
|
| 869 |
+
"step": 9700
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.7848790645522986,
|
| 873 |
+
"grad_norm": 0.27292346954345703,
|
| 874 |
+
"learning_rate": 2.391207617691555e-05,
|
| 875 |
+
"loss": 0.0941,
|
| 876 |
+
"step": 9800
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 0.7928880345987506,
|
| 880 |
+
"grad_norm": 0.24669557809829712,
|
| 881 |
+
"learning_rate": 2.3022158939218656e-05,
|
| 882 |
+
"loss": 0.0925,
|
| 883 |
+
"step": 9900
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 0.8008970046452026,
|
| 887 |
+
"grad_norm": 0.2944881021976471,
|
| 888 |
+
"learning_rate": 2.213224170152176e-05,
|
| 889 |
+
"loss": 0.0945,
|
| 890 |
+
"step": 10000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.8089059746916547,
|
| 894 |
+
"grad_norm": 0.09000946581363678,
|
| 895 |
+
"learning_rate": 2.1242324463824865e-05,
|
| 896 |
+
"loss": 0.0857,
|
| 897 |
+
"step": 10100
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.8169149447381067,
|
| 901 |
+
"grad_norm": 0.27225881814956665,
|
| 902 |
+
"learning_rate": 2.035240722612797e-05,
|
| 903 |
+
"loss": 0.0854,
|
| 904 |
+
"step": 10200
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 0.8249239147845587,
|
| 908 |
+
"grad_norm": 0.19183236360549927,
|
| 909 |
+
"learning_rate": 1.9462489988431076e-05,
|
| 910 |
+
"loss": 0.0876,
|
| 911 |
+
"step": 10300
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 0.8329328848310107,
|
| 915 |
+
"grad_norm": 0.4199092984199524,
|
| 916 |
+
"learning_rate": 1.8572572750734184e-05,
|
| 917 |
+
"loss": 0.0896,
|
| 918 |
+
"step": 10400
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 0.8409418548774628,
|
| 922 |
+
"grad_norm": 0.16657961905002594,
|
| 923 |
+
"learning_rate": 1.7682655513037288e-05,
|
| 924 |
+
"loss": 0.0832,
|
| 925 |
+
"step": 10500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 0.8409418548774628,
|
| 929 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.07417957921898752,
|
| 930 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.08394235372543335,
|
| 931 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 237.0276,
|
| 932 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.109,
|
| 933 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.266,
|
| 934 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.1648900933453779,
|
| 935 |
+
"step": 10500
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.8409418548774628,
|
| 939 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.10516655960356061,
|
| 940 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.28152379393577576,
|
| 941 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 39.2346,
|
| 942 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 12.438,
|
| 943 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.555,
|
| 944 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.2734405649274225,
|
| 945 |
+
"step": 10500
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.8409418548774628,
|
| 949 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5044642857142857,
|
| 950 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5193919539451599,
|
| 951 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 26.7813,
|
| 952 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 3.024,
|
| 953 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.411,
|
| 954 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.41278439869989164,
|
| 955 |
+
"step": 10500
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.8489508249239148,
|
| 959 |
+
"grad_norm": 0.25866174697875977,
|
| 960 |
+
"learning_rate": 1.6792738275340396e-05,
|
| 961 |
+
"loss": 0.0853,
|
| 962 |
+
"step": 10600
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.8569597949703668,
|
| 966 |
+
"grad_norm": 0.28780218958854675,
|
| 967 |
+
"learning_rate": 1.59028210376435e-05,
|
| 968 |
+
"loss": 0.0899,
|
| 969 |
+
"step": 10700
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.8649687650168189,
|
| 973 |
+
"grad_norm": 0.3092251121997833,
|
| 974 |
+
"learning_rate": 1.5012903799946607e-05,
|
| 975 |
+
"loss": 0.0949,
|
| 976 |
+
"step": 10800
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.8729777350632708,
|
| 980 |
+
"grad_norm": 0.1439545601606369,
|
| 981 |
+
"learning_rate": 1.4122986562249713e-05,
|
| 982 |
+
"loss": 0.0846,
|
| 983 |
+
"step": 10900
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.8809867051097229,
|
| 987 |
+
"grad_norm": 0.49992987513542175,
|
| 988 |
+
"learning_rate": 1.3233069324552816e-05,
|
| 989 |
+
"loss": 0.0878,
|
| 990 |
+
"step": 11000
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.888995675156175,
|
| 994 |
+
"grad_norm": 0.5189112424850464,
|
| 995 |
+
"learning_rate": 1.2343152086855923e-05,
|
| 996 |
+
"loss": 0.088,
|
| 997 |
+
"step": 11100
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.8970046452026269,
|
| 1001 |
+
"grad_norm": 0.22200240194797516,
|
| 1002 |
+
"learning_rate": 1.1453234849159029e-05,
|
| 1003 |
+
"loss": 0.0793,
|
| 1004 |
+
"step": 11200
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.905013615249079,
|
| 1008 |
+
"grad_norm": 0.8681122660636902,
|
| 1009 |
+
"learning_rate": 1.0563317611462133e-05,
|
| 1010 |
+
"loss": 0.0897,
|
| 1011 |
+
"step": 11300
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.913022585295531,
|
| 1015 |
+
"grad_norm": 0.1663997769355774,
|
| 1016 |
+
"learning_rate": 9.673400373765241e-06,
|
| 1017 |
+
"loss": 0.0865,
|
| 1018 |
+
"step": 11400
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.921031555341983,
|
| 1022 |
+
"grad_norm": 0.3327358067035675,
|
| 1023 |
+
"learning_rate": 8.783483136068347e-06,
|
| 1024 |
+
"loss": 0.0887,
|
| 1025 |
+
"step": 11500
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.9290405253884351,
|
| 1029 |
+
"grad_norm": 0.23733209073543549,
|
| 1030 |
+
"learning_rate": 7.893565898371453e-06,
|
| 1031 |
+
"loss": 0.0868,
|
| 1032 |
+
"step": 11600
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.937049495434887,
|
| 1036 |
+
"grad_norm": 0.21292167901992798,
|
| 1037 |
+
"learning_rate": 7.0036486606745585e-06,
|
| 1038 |
+
"loss": 0.0896,
|
| 1039 |
+
"step": 11700
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.9450584654813391,
|
| 1043 |
+
"grad_norm": 0.21337737143039703,
|
| 1044 |
+
"learning_rate": 6.1137314229776635e-06,
|
| 1045 |
+
"loss": 0.0847,
|
| 1046 |
+
"step": 11800
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.9530674355277912,
|
| 1050 |
+
"grad_norm": 0.23139464855194092,
|
| 1051 |
+
"learning_rate": 5.223814185280769e-06,
|
| 1052 |
+
"loss": 0.082,
|
| 1053 |
+
"step": 11900
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.9610764055742431,
|
| 1057 |
+
"grad_norm": 0.4367704391479492,
|
| 1058 |
+
"learning_rate": 4.333896947583875e-06,
|
| 1059 |
+
"loss": 0.0874,
|
| 1060 |
+
"step": 12000
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.9610764055742431,
|
| 1064 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__cer": 0.07097326367155066,
|
| 1065 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__loss": 0.08283345401287079,
|
| 1066 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__runtime": 236.9213,
|
| 1067 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__samples_per_second": 2.11,
|
| 1068 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__steps_per_second": 0.266,
|
| 1069 |
+
"eval_oza75_bm-text-normalization__cbr__dev__500__wer": 0.16193917494730503,
|
| 1070 |
+
"step": 12000
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.9610764055742431,
|
| 1074 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__cer": 0.1025052766816555,
|
| 1075 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__loss": 0.27890169620513916,
|
| 1076 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__runtime": 41.0822,
|
| 1077 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__samples_per_second": 11.879,
|
| 1078 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__steps_per_second": 1.485,
|
| 1079 |
+
"eval_djelia_bm-text-normalization__bamadaba__dev__500__wer": 0.2742251863475873,
|
| 1080 |
+
"step": 12000
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.9610764055742431,
|
| 1084 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_cer": 0.5032894736842105,
|
| 1085 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_loss": 0.5151007771492004,
|
| 1086 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_runtime": 27.495,
|
| 1087 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_samples_per_second": 2.946,
|
| 1088 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_steps_per_second": 0.4,
|
| 1089 |
+
"eval_djelia_bm-text-normalization-benchmark__level-1__test_wer": 0.4106175514626219,
|
| 1090 |
+
"step": 12000
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.9690853756206952,
|
| 1094 |
+
"grad_norm": 0.18608327209949493,
|
| 1095 |
+
"learning_rate": 3.4439797098869802e-06,
|
| 1096 |
+
"loss": 0.0888,
|
| 1097 |
+
"step": 12100
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 0.9770943456671473,
|
| 1101 |
+
"grad_norm": 0.3838459253311157,
|
| 1102 |
+
"learning_rate": 2.5540624721900865e-06,
|
| 1103 |
+
"loss": 0.0834,
|
| 1104 |
+
"step": 12200
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.9851033157135992,
|
| 1108 |
+
"grad_norm": 0.32492393255233765,
|
| 1109 |
+
"learning_rate": 1.6641452344931922e-06,
|
| 1110 |
+
"loss": 0.0862,
|
| 1111 |
+
"step": 12300
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 0.9931122857600513,
|
| 1115 |
+
"grad_norm": 0.3034497797489166,
|
| 1116 |
+
"learning_rate": 7.74227996796298e-07,
|
| 1117 |
+
"loss": 0.0879,
|
| 1118 |
+
"step": 12400
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 1.0,
|
| 1122 |
+
"step": 12486,
|
| 1123 |
+
"total_flos": 1.1016496820713882e+17,
|
| 1124 |
+
"train_loss": 0.1502396887776639,
|
| 1125 |
+
"train_runtime": 5207.863,
|
| 1126 |
+
"train_samples_per_second": 19.179,
|
| 1127 |
+
"train_steps_per_second": 2.398
|
| 1128 |
+
}
|
| 1129 |
+
],
|
| 1130 |
+
"logging_steps": 100,
|
| 1131 |
+
"max_steps": 12486,
|
| 1132 |
+
"num_input_tokens_seen": 0,
|
| 1133 |
+
"num_train_epochs": 1,
|
| 1134 |
+
"save_steps": 1500,
|
| 1135 |
+
"stateful_callbacks": {
|
| 1136 |
+
"TrainerControl": {
|
| 1137 |
+
"args": {
|
| 1138 |
+
"should_epoch_stop": false,
|
| 1139 |
+
"should_evaluate": false,
|
| 1140 |
+
"should_log": false,
|
| 1141 |
+
"should_save": true,
|
| 1142 |
+
"should_training_stop": true
|
| 1143 |
+
},
|
| 1144 |
+
"attributes": {}
|
| 1145 |
+
}
|
| 1146 |
+
},
|
| 1147 |
+
"total_flos": 1.1016496820713882e+17,
|
| 1148 |
+
"train_batch_size": 8,
|
| 1149 |
+
"trial_name": null,
|
| 1150 |
+
"trial_params": null
|
| 1151 |
+
}
|