kimbochen commited on
Commit
280c652
1 Parent(s): 8a813aa

Training in progress, step 400

Browse files
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "openai/whisper-small",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
  "architectures": [
 
1
  {
2
+ "_name_or_path": "kimbochen/whisper-small-zh-tw",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
  "architectures": [
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:478a3830dcece05bf882aeb9d15c5f2cb5a7ec9456d33f7a23642664f73e08a1
3
  size 967102601
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f1048f1d52866be74ce2c8fc032b3a2b6cbf7f987c43ad4ba97882d7a06806e
3
  size 967102601
run.sh CHANGED
@@ -1,6 +1,6 @@
1
  python run_speech_recognition_seq2seq_streaming.py \
2
  --output_dir="./" \
3
- --model_name_or_path="openai/whisper-small" \
4
  --model_index_name="Whisper Small Traditional Chinese" \
5
  --dataset_name="mozilla-foundation/common_voice_11_0" \
6
  --dataset_config_name="zh-TW" \
@@ -10,13 +10,13 @@ python run_speech_recognition_seq2seq_streaming.py \
10
  --learning_rate="1e-5" \
11
  --per_device_train_batch_size="64" \
12
  --per_device_eval_batch_size="32" \
13
- --max_steps="1000" \
14
- --warmup_steps="800" \
15
  --logging_steps="25" \
16
  --evaluation_strategy="steps" \
17
- --eval_steps="200" \
18
  --save_strategy="steps" \
19
- --save_steps="200" \
20
  --generation_max_length="225" \
21
  --length_column_name="input_length" \
22
  --max_duration_in_seconds="30" \
 
1
  python run_speech_recognition_seq2seq_streaming.py \
2
  --output_dir="./" \
3
+ --model_name_or_path="kimbochen/whisper-small-zh-tw" \
4
  --model_index_name="Whisper Small Traditional Chinese" \
5
  --dataset_name="mozilla-foundation/common_voice_11_0" \
6
  --dataset_config_name="zh-TW" \
 
10
  --learning_rate="1e-5" \
11
  --per_device_train_batch_size="64" \
12
  --per_device_eval_batch_size="32" \
13
+ --max_steps="2000" \
14
+ --warmup_steps="1200" \
15
  --logging_steps="25" \
16
  --evaluation_strategy="steps" \
17
+ --eval_steps="400" \
18
  --save_strategy="steps" \
19
+ --save_steps="400" \
20
  --generation_max_length="225" \
21
  --length_column_name="input_length" \
22
  --max_duration_in_seconds="30" \
runs/Dec14_01-26-33_129-213-23-45/1670981229.51901/events.out.tfevents.1670981229.129-213-23-45.2845517.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:624c5a48c93b287bdf5cba11b1a8b85a19514ff62d4def8b0e6ddc56eed628e7
3
+ size 5869
runs/Dec14_01-26-33_129-213-23-45/events.out.tfevents.1670981229.129-213-23-45.2845517.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3b36c39531d78761a66dd1e9844edb63f58bb4a193af20b2dcf5a5218ba40da
3
+ size 7089
tokenizer_config.json CHANGED
@@ -19,7 +19,7 @@
19
  },
20
  "errors": "replace",
21
  "model_max_length": 1024,
22
- "name_or_path": "openai/whisper-small",
23
  "pad_token": null,
24
  "processor_class": "WhisperProcessor",
25
  "return_attention_mask": false,
 
19
  },
20
  "errors": "replace",
21
  "model_max_length": 1024,
22
+ "name_or_path": "kimbochen/whisper-small-zh-tw",
23
  "pad_token": null,
24
  "processor_class": "WhisperProcessor",
25
  "return_attention_mask": false,
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a25dac4d63073c01620ce6bec3c37e5ff05a5b0b236e2dc7fc2530c8f0726c0c
3
  size 3579
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50004a441a60d2899a3d26a54f48491d08480f13b412fc7bd5ee8fa6c1f914c3
3
  size 3579