patrickvonplaten
commited on
Commit
•
07c73c4
1
Parent(s):
c1433ec
up
Browse files- all_results.json +5 -5
- config.json +2 -2
- create_model.py +2 -1
- emissions.csv +1 -0
- pytorch_model.bin +2 -2
- run_librispeech.sh +3 -2
- runs/Dec23_10-49-11_ip-172-31-1-149/1640256589.3304777/events.out.tfevents.1640256589.ip-172-31-1-149.645478.1 +3 -0
- runs/Dec23_10-49-11_ip-172-31-1-149/events.out.tfevents.1640256589.ip-172-31-1-149.645478.0 +3 -0
- runs/Dec23_10-57-06_ip-172-31-1-149/1640257063.6087983/events.out.tfevents.1640257063.ip-172-31-1-149.714399.1 +3 -0
- runs/Dec23_10-57-06_ip-172-31-1-149/events.out.tfevents.1640257063.ip-172-31-1-149.714399.0 +3 -0
- train_results.json +5 -5
- trainer_state.json +0 -0
- training_args.bin +1 -1
all_results.json
CHANGED
@@ -1,14 +1,14 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
"eval_loss": 23.11814308166504,
|
4 |
"eval_runtime": 563.3925,
|
5 |
"eval_samples": 2642,
|
6 |
"eval_samples_per_second": 4.689,
|
7 |
"eval_steps_per_second": 0.147,
|
8 |
"eval_wer": 1.0,
|
9 |
-
"train_loss":
|
10 |
-
"train_runtime":
|
11 |
"train_samples": 28538,
|
12 |
-
"train_samples_per_second":
|
13 |
-
"train_steps_per_second": 0.
|
14 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 5.0,
|
3 |
"eval_loss": 23.11814308166504,
|
4 |
"eval_runtime": 563.3925,
|
5 |
"eval_samples": 2642,
|
6 |
"eval_samples_per_second": 4.689,
|
7 |
"eval_steps_per_second": 0.147,
|
8 |
"eval_wer": 1.0,
|
9 |
+
"train_loss": 1.844849973621323,
|
10 |
+
"train_runtime": 5604.6492,
|
11 |
"train_samples": 28538,
|
12 |
+
"train_samples_per_second": 25.459,
|
13 |
+
"train_steps_per_second": 0.398
|
14 |
}
|
config.json
CHANGED
@@ -173,7 +173,7 @@
|
|
173 |
"feat_extract_norm": "layer",
|
174 |
"feat_proj_dropout": 0.0,
|
175 |
"feat_quantizer_dropout": 0.0,
|
176 |
-
"final_dropout": 0.
|
177 |
"finetuning_task": null,
|
178 |
"forced_bos_token_id": null,
|
179 |
"forced_eos_token_id": null,
|
@@ -202,7 +202,7 @@
|
|
202 |
"mask_feature_prob": 0.0,
|
203 |
"mask_time_length": 10,
|
204 |
"mask_time_min_masks": 2,
|
205 |
-
"mask_time_prob": 0.
|
206 |
"max_length": 20,
|
207 |
"min_length": 0,
|
208 |
"model_type": "wav2vec2",
|
|
|
173 |
"feat_extract_norm": "layer",
|
174 |
"feat_proj_dropout": 0.0,
|
175 |
"feat_quantizer_dropout": 0.0,
|
176 |
+
"final_dropout": 0.0,
|
177 |
"finetuning_task": null,
|
178 |
"forced_bos_token_id": null,
|
179 |
"forced_eos_token_id": null,
|
|
|
202 |
"mask_feature_prob": 0.0,
|
203 |
"mask_time_length": 10,
|
204 |
"mask_time_min_masks": 2,
|
205 |
+
"mask_time_prob": 0.1,
|
206 |
"max_length": 20,
|
207 |
"min_length": 0,
|
208 |
"model_type": "wav2vec2",
|
create_model.py
CHANGED
@@ -8,7 +8,8 @@ decoder_id = "facebook/bart-large"
|
|
8 |
|
9 |
model = SpeechEncoderDecoderModel.from_encoder_decoder_pretrained(encoder_id, decoder_id, encoder_add_adapter=True)
|
10 |
model.config.encoder.feat_proj_dropout = 0.0
|
11 |
-
model.config.encoder.
|
|
|
12 |
model.config.decoder_start_token_id = model.decoder.config.bos_token_id
|
13 |
model.config.pad_token_id = model.decoder.config.pad_token_id
|
14 |
model.config.eos_token_id = model.decoder.config.eos_token_id
|
|
|
8 |
|
9 |
model = SpeechEncoderDecoderModel.from_encoder_decoder_pretrained(encoder_id, decoder_id, encoder_add_adapter=True)
|
10 |
model.config.encoder.feat_proj_dropout = 0.0
|
11 |
+
model.config.encoder.final_dropout = 0.0
|
12 |
+
model.config.encoder.mask_time_prob = 0.1
|
13 |
model.config.decoder_start_token_id = model.decoder.config.bos_token_id
|
14 |
model.config.pad_token_id = model.decoder.config.pad_token_id
|
15 |
model.config.eos_token_id = model.decoder.config.eos_token_id
|
emissions.csv
CHANGED
@@ -1,2 +1,3 @@
|
|
1 |
timestamp,experiment_id,project_name,duration,emissions,energy_consumed,country_name,country_iso_code,region,on_cloud,cloud_provider,cloud_region
|
2 |
2021-12-23T04:13:13,7231458a-1d3e-45ac-82fc-95d20ece134c,codecarbon,28532.068876504898,4.1430815744945955,11.264495852350722,USA,USA,nan,Y,aws,us-east-1
|
|
|
|
1 |
timestamp,experiment_id,project_name,duration,emissions,energy_consumed,country_name,country_iso_code,region,on_cloud,cloud_provider,cloud_region
|
2 |
2021-12-23T04:13:13,7231458a-1d3e-45ac-82fc-95d20ece134c,codecarbon,28532.068876504898,4.1430815744945955,11.264495852350722,USA,USA,nan,Y,aws,us-east-1
|
3 |
+
2021-12-23T12:31:06,beb31896-3df5-43c8-9163-a3d02456a623,codecarbon,5600.265459775925,0.8146651355314377,2.2149677420648115,USA,USA,nan,Y,aws,us-east-1
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c3128180dc6f3faf1bbf77470f8f5f25284d19d333976e4b0bc855ef517990f5
|
3 |
+
size 2278381245
|
run_librispeech.sh
CHANGED
@@ -11,10 +11,10 @@ python -m torch.distributed.launch \
|
|
11 |
--preprocessing_num_workers="16" \
|
12 |
--length_column_name="input_length" \
|
13 |
--overwrite_output_dir \
|
14 |
-
--num_train_epochs="
|
15 |
--per_device_train_batch_size="4" \
|
16 |
--per_device_eval_batch_size="4" \
|
17 |
-
--gradient_accumulation_steps="
|
18 |
--generation_max_length="40" \
|
19 |
--generation_num_beams="1" \
|
20 |
--learning_rate="3e-4" \
|
@@ -30,4 +30,5 @@ python -m torch.distributed.launch \
|
|
30 |
--fp16 \
|
31 |
--group_by_length \
|
32 |
--predict_with_generate \
|
|
|
33 |
--do_eval --do_train
|
|
|
11 |
--preprocessing_num_workers="16" \
|
12 |
--length_column_name="input_length" \
|
13 |
--overwrite_output_dir \
|
14 |
+
--num_train_epochs="5" \
|
15 |
--per_device_train_batch_size="4" \
|
16 |
--per_device_eval_batch_size="4" \
|
17 |
+
--gradient_accumulation_steps="2" \
|
18 |
--generation_max_length="40" \
|
19 |
--generation_num_beams="1" \
|
20 |
--learning_rate="3e-4" \
|
|
|
30 |
--fp16 \
|
31 |
--group_by_length \
|
32 |
--predict_with_generate \
|
33 |
+
--do_lower_case \
|
34 |
--do_eval --do_train
|
runs/Dec23_10-49-11_ip-172-31-1-149/1640256589.3304777/events.out.tfevents.1640256589.ip-172-31-1-149.645478.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fa39857b3345487112d203460efc1683d1d5c47ad9c17ad4c726875a77c8eab1
|
3 |
+
size 4957
|
runs/Dec23_10-49-11_ip-172-31-1-149/events.out.tfevents.1640256589.ip-172-31-1-149.645478.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6cf1064c7262f6db0ec0a26ffb6db7f278d6e43449c768502b4d968cafabbe0c
|
3 |
+
size 11608
|
runs/Dec23_10-57-06_ip-172-31-1-149/1640257063.6087983/events.out.tfevents.1640257063.ip-172-31-1-149.714399.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f9a8004194c841a35c30ccfa2b691f69fce2812780667915857670bd098314d3
|
3 |
+
size 4957
|
runs/Dec23_10-57-06_ip-172-31-1-149/events.out.tfevents.1640257063.ip-172-31-1-149.714399.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9414a120df6d371de49275ec6a034d69e8937014b66a4302c635687686668c28
|
3 |
+
size 360963
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 28538,
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 5.0,
|
3 |
+
"train_loss": 1.844849973621323,
|
4 |
+
"train_runtime": 5604.6492,
|
5 |
"train_samples": 28538,
|
6 |
+
"train_samples_per_second": 25.459,
|
7 |
+
"train_steps_per_second": 0.398
|
8 |
}
|
trainer_state.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3055
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0ebcadf8a557cc2a21ac9306cd8a4038db6cdcad1cd7a661d41c3de1913899af
|
3 |
size 3055
|