Rolv-Arild commited on
Commit
c317412
β€’
1 Parent(s): 586aeaf

Training in progress, step 1500

Browse files
ex3/vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"a": 1, "b": 2, "c": 3, "d": 4, "e": 5, "f": 6, "g": 7, "h": 8, "i": 9, "j": 10, "k": 11, "l": 12, "m": 13, "n": 14, "o": 15, "p": 16, "q": 17, "r": 18, "s": 19, "t": 20, "u": 21, "v": 22, "w": 23, "x": 24, "y": 25, "z": 26, "\u00e5": 27, "\u00e6": 28, "\u00f8": 29, "|": 0, "[UNK]": 30, "[PAD]": 31}
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1db06c6f208919999109fb54084bd07850ef535e6dbd5edad7741b1b78a55c7b
3
  size 1262063089
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6edf3b6d111f124f397bb09783b024609d2532d55c30be41df0e5b870da8916
3
  size 1262063089
run.sh CHANGED
@@ -8,7 +8,7 @@ python run_speech_recognition_ctc.py \
8
  --per_device_train_batch_size="16" \
9
  --per_device_eval_batch_size="16" \
10
  --gradient_accumulation_steps="4" \
11
- --learning_rate="3e-5" \
12
  --warmup_steps="500" \
13
  --length_column_name="input_length" \
14
  --evaluation_strategy="steps" \
 
8
  --per_device_train_batch_size="16" \
9
  --per_device_eval_batch_size="16" \
10
  --gradient_accumulation_steps="4" \
11
+ --learning_rate="7.5e-5" \
12
  --warmup_steps="500" \
13
  --length_column_name="input_length" \
14
  --evaluation_strategy="steps" \
run_ex1.sh ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ python run_speech_recognition_ctc.py \
2
+ --dataset_name="NbAiLab/NPSC" \
3
+ --model_name_or_path="facebook/wav2vec2-xls-r-300m" \
4
+ --dataset_config_name="16K_mp3" \
5
+ --output_dir="./" \
6
+ --overwrite_output_dir \
7
+ --num_train_epochs="20" \
8
+ --per_device_train_batch_size="16" \
9
+ --per_device_eval_batch_size="16" \
10
+ --gradient_accumulation_steps="4" \
11
+ --learning_rate="3e-5" \
12
+ --warmup_steps="500" \
13
+ --length_column_name="input_length" \
14
+ --evaluation_strategy="steps" \
15
+ --text_column_name="text" \
16
+ --save_steps="250" \
17
+ --eval_steps="250" \
18
+ --logging_steps="50" \
19
+ --layerdrop="0.0" \
20
+ --activation_dropout="0.1" \
21
+ --save_total_limit="3" \
22
+ --freeze_feature_encoder \
23
+ --feat_proj_dropout="0.0" \
24
+ --mask_time_prob="0.75" \
25
+ --mask_time_length="10" \
26
+ --mask_feature_prob="0.25" \
27
+ --mask_feature_length="64" \
28
+ --chars_to_ignore , ? . ! \- \; \: \" β€œ % β€˜ ” οΏ½ β€” ’ … – _ \\ + \# / \
29
+ --gradient_checkpointing \
30
+ --use_auth_token \
31
+ --fp16 \
32
+ --group_by_length \
33
+ --do_train --do_eval \
34
+ --push_to_hub \
35
+ --preprocessing_num_workers="8"
36
+
run_ex3.sh ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ python run_speech_recognition_ctc.py \
2
+ --dataset_name="NbAiLab/NPSC" \
3
+ --model_name_or_path="facebook/wav2vec2-xls-r-300m" \
4
+ --dataset_config_name="16K_mp3" \
5
+ --output_dir="./ex3" \
6
+ --overwrite_output_dir \
7
+ --num_train_epochs="20" \
8
+ --per_device_train_batch_size="16" \
9
+ --per_device_eval_batch_size="16" \
10
+ --gradient_accumulation_steps="4" \
11
+ --learning_rate="7.5e-5" \
12
+ --warmup_steps="500" \
13
+ --length_column_name="input_length" \
14
+ --evaluation_strategy="steps" \
15
+ --text_column_name="text" \
16
+ --save_steps="250" \
17
+ --eval_steps="250" \
18
+ --logging_steps="50" \
19
+ --layerdrop="0.1" \
20
+ --activation_dropout="0.1" \
21
+ --save_total_limit="3" \
22
+ --freeze_feature_encoder \
23
+ --feat_proj_dropout="0.0" \
24
+ --mask_time_prob="0.75" \
25
+ --mask_time_length="10" \
26
+ --mask_feature_prob="0.25" \
27
+ --mask_feature_length="64" \
28
+ --chars_to_ignore , ? . ! \- \; \: \" β€œ % β€˜ ” οΏ½ β€” ’ … – _ \\ + \# / \
29
+ --gradient_checkpointing \
30
+ --use_auth_token \
31
+ --fp16 \
32
+ --group_by_length \
33
+ --do_train --do_eval \
34
+ --push_to_hub \
35
+ --preprocessing_num_workers="8"
36
+
runs/Jan31_11-10-54_ficino/events.out.tfevents.1643623929.ficino.269307.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:60f37df178c0c2b78acfd6de36eefa08f745e7ff73bedadfd56a1b79d1977423
3
- size 10155
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72a77bf0403a71bedf47a4a64b53031e1d7f37627570e4ea9710a52d44eed7ca
3
+ size 11258