Hugging Face's logo

sanchit-gandhi
/
parler-tts-600M-cross-attention

Text2Text Generation
Transformers
parler_tts
Inference Endpoints
Model card Files Community
sanchit-gandhi HF staff commited on
Commit
866256b
•
1 Parent(s): 0fcd0cb

Saving train state of step 90000

Browse files
Files changed (16) hide show
  1. checkpoint-90000-epoch-6/optimizer.bin +3 -0
  2. checkpoint-90000-epoch-6/pytorch_model.bin +3 -0
  3. checkpoint-90000-epoch-6/random_states_0.pkl +3 -0
  4. checkpoint-90000-epoch-6/random_states_1.pkl +3 -0
  5. checkpoint-90000-epoch-6/random_states_2.pkl +3 -0
  6. checkpoint-90000-epoch-6/random_states_3.pkl +3 -0
  7. checkpoint-90000-epoch-6/random_states_4.pkl +3 -0
  8. checkpoint-90000-epoch-6/random_states_5.pkl +3 -0
  9. checkpoint-90000-epoch-6/random_states_6.pkl +3 -0
  10. checkpoint-90000-epoch-6/random_states_7.pkl +3 -0
  11. checkpoint-90000-epoch-6/scheduler.bin +3 -0
  12. parler_tts/__pycache__/configuration_parler_tts.cpython-311.pyc +0 -0
  13. parler_tts/__pycache__/modeling_parler_tts.cpython-311.pyc +0 -0
  14. slurm_job.slurm +0 -1
  15. starting_point_0.01.json +1 -1
  16. training/eval.py +3 -2
checkpoint-90000-epoch-6/optimizer.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1eb3ae26d148706a8971642f2f08005ca8aff8f0191c6dbfe9927263ce4356da
3
+ size 3652769047
checkpoint-90000-epoch-6/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c1d67cfacbeb89775d04e0b05a9f89ac4d7c3279add6b03368bff8dc957f1af
3
+ size 2605239710
checkpoint-90000-epoch-6/random_states_0.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a5b64952dc0f58fc23466f35d49986a91e360b6b179ee8e95e03490b1fb9db5
3
+ size 16100
checkpoint-90000-epoch-6/random_states_1.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b748c0691053cd51f3a3723fc79fcc0df47bba24fddf1f0190d88a66b94ab03d
3
+ size 16164
checkpoint-90000-epoch-6/random_states_2.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2588b7c3d53a410f3bc598dcd6fd3e61d9263bd35aee07d3676090051dcf4c15
3
+ size 16164
checkpoint-90000-epoch-6/random_states_3.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af6e7950307ad612a3d951930af7474ffe99794b7ab53b444e83bfc0d16dc984
3
+ size 16100
checkpoint-90000-epoch-6/random_states_4.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cfa3f0172fe672362b6be1f8c0e31fdc1501e138e7107c5a80572da1fc58c05
3
+ size 16100
checkpoint-90000-epoch-6/random_states_5.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8b34c63cdb39558f32d242a7788851435883e67d09dd98b59a0730cbbffc784
3
+ size 16100
checkpoint-90000-epoch-6/random_states_6.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20bd8150c333c8f7440b2756d1cc255e3b057232aa06122612efc8630f521ea7
3
+ size 16100
checkpoint-90000-epoch-6/random_states_7.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e54b1d060c27f28a6629dad2fd6ec072e8682b4a6efe184614a03c732056bee7
3
+ size 16100
checkpoint-90000-epoch-6/scheduler.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:450723525f8289e0c84483655a1f71b4aacf9b3f252565f47a390fb44ebcb50c
3
+ size 1000
parler_tts/__pycache__/configuration_parler_tts.cpython-311.pyc CHANGED
Binary files a/parler_tts/__pycache__/configuration_parler_tts.cpython-311.pyc and b/parler_tts/__pycache__/configuration_parler_tts.cpython-311.pyc differ
 
parler_tts/__pycache__/modeling_parler_tts.cpython-311.pyc CHANGED
Binary files a/parler_tts/__pycache__/modeling_parler_tts.cpython-311.pyc and b/parler_tts/__pycache__/modeling_parler_tts.cpython-311.pyc differ
 
slurm_job.slurm CHANGED
@@ -3,7 +3,6 @@
3
  #SBATCH --nodes=1
4
  # set 48h for job wall time limit
5
  #SBATCH --time=48:00:00
6
- #SBATCH --requeue
7
  #SBATCH --ntasks-per-node=1 # crucial - only 1 task per dist per node!
8
  #SBATCH --cpus-per-task=32
9
  #SBATCH --gres=gpu:8
 
3
  #SBATCH --nodes=1
4
  # set 48h for job wall time limit
5
  #SBATCH --time=48:00:00
 
6
  #SBATCH --ntasks-per-node=1 # crucial - only 1 task per dist per node!
7
  #SBATCH --cpus-per-task=32
8
  #SBATCH --gres=gpu:8
starting_point_0.01.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "model_name_or_path": "parler-tts/parler-tts-untrained-600M-cross-attention",
3
- "save_to_disk": "/scratch/tmp_dataset_audio/",
4
  "temporary_save_to_disk": "/scratch/tmp_dataset_audio/",
5
  "push_to_hub": true,
6
 
 
1
  {
2
  "model_name_or_path": "parler-tts/parler-tts-untrained-600M-cross-attention",
3
+ "save_to_disk": "/fsx/yoach/tmp/artefacts/10k_hours_processed_punctuated/",
4
  "temporary_save_to_disk": "/scratch/tmp_dataset_audio/",
5
  "push_to_hub": true,
6
 
training/eval.py CHANGED
@@ -47,12 +47,13 @@ def wer(asr_model_name_or_path, prompts, audios, device, per_device_eval_batch_s
47
  normalized_references = []
48
 
49
  for pred, ref in zip(transcriptions, prompts):
50
- normalizer = english_normalizer if hasattr(pred, "language") and pred["language"] == "english" else basic_normalizer
 
51
  norm_ref = normalizer(ref)
52
  if len(norm_ref) > 0:
53
  norm_pred = normalizer(pred["text"])
54
  normalized_predictions.append(norm_pred)
55
- normalized_references.append(norm_pred)
56
 
57
  word_error = 100 * metric.compute(predictions=normalized_predictions, references=normalized_references)
58
 
 
47
  normalized_references = []
48
 
49
  for pred, ref in zip(transcriptions, prompts):
50
+ normalizer = english_normalizer
51
+
52
  norm_ref = normalizer(ref)
53
  if len(norm_ref) > 0:
54
  norm_pred = normalizer(pred["text"])
55
  normalized_predictions.append(norm_pred)
56
+ normalized_references.append(norm_ref)
57
 
58
  word_error = 100 * metric.compute(predictions=normalized_predictions, references=normalized_references)
59