Hugging Face's logo

ihanif
/
whisper-medium-urdu

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Urdu
whisper
hf-asr-leaderboard
urdu
whisper-event
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
2
ihanif commited on
Commit
49b3f35
•
1 Parent(s): bee5848

Training in progress, step 300

Browse files
Files changed (8) hide show
  1. all_results.json +7 -0
  2. pytorch_model.bin +1 -1
  3. run.sh +2 -2
  4. runs/Dec11_14-58-14_150-136-41-252/1670770901.4028096/events.out.tfevents.1670770901.150-136-41-252.369844.1 +3 -0
  5. runs/Dec11_14-58-14_150-136-41-252/events.out.tfevents.1670770901.150-136-41-252.369844.0 +3 -0
  6. train_results.json +7 -0
  7. trainer_state.json +163 -0
  8. training_args.bin +1 -1
all_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.97,
3
+ "train_loss": 0.4665303599834442,
4
+ "train_runtime": 6285.9399,
5
+ "train_samples_per_second": 1.018,
6
+ "train_steps_per_second": 0.032
7
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5eeaa4b4b2e7dc750a86f0904797221f08de5028239fc72f25bd276a772000c1
3
  size 3055754841
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc43fe323b4439210889594b2ed3475e474ed2a8ccd3f5663693c5ac4386193a
3
  size 3055754841
run.sh CHANGED
@@ -6,7 +6,7 @@ python run_speech_recognition_seq2seq_streaming.py \
6
  --train_split_name="train+validation" \
7
  --eval_split_name="test" \
8
  --model_index_name="Whisper Medium Urdu" \
9
- --max_steps="200" \
10
  --output_dir="./" \
11
  --per_device_train_batch_size="32" \
12
  --per_device_eval_batch_size="16" \
@@ -29,7 +29,7 @@ python run_speech_recognition_seq2seq_streaming.py \
29
  --load_best_model_at_end="True" \
30
  --gradient_checkpointing \
31
  --fp16 \
32
- --overwrite_output_dir="True" \
33
  --do_train \
34
  --do_eval \
35
  --predict_with_generate \
 
6
  --train_split_name="train+validation" \
7
  --eval_split_name="test" \
8
  --model_index_name="Whisper Medium Urdu" \
9
+ --max_steps="300" \
10
  --output_dir="./" \
11
  --per_device_train_batch_size="32" \
12
  --per_device_eval_batch_size="16" \
 
29
  --load_best_model_at_end="True" \
30
  --gradient_checkpointing \
31
  --fp16 \
32
+ --overwrite_output_dir="False" \
33
  --do_train \
34
  --do_eval \
35
  --predict_with_generate \
runs/Dec11_14-58-14_150-136-41-252/1670770901.4028096/events.out.tfevents.1670770901.150-136-41-252.369844.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b11d6294211e96633db9f4c117fa959e703a4f521759e73851dfad95713163ef
3
+ size 5870
runs/Dec11_14-58-14_150-136-41-252/events.out.tfevents.1670770901.150-136-41-252.369844.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:58962c526f8a25aea9f9802325eef5841215cdf6b2cd5db2300fc2f2c25df99a
3
+ size 6154
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.97,
3
+ "train_loss": 0.4665303599834442,
4
+ "train_runtime": 6285.9399,
5
+ "train_samples_per_second": 1.018,
6
+ "train_steps_per_second": 0.032
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,163 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 26.980130911344357,
3
+ "best_model_checkpoint": "./checkpoint-200",
4
+ "epoch": 0.966183574879227,
5
+ "global_step": 200,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.05,
12
+ "learning_rate": 2.0000000000000003e-06,
13
+ "loss": 0.9955,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.1,
18
+ "learning_rate": 4.5e-06,
19
+ "loss": 0.7842,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.14,
24
+ "learning_rate": 7e-06,
25
+ "loss": 0.533,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.19,
30
+ "learning_rate": 9.5e-06,
31
+ "loss": 0.5079,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.24,
36
+ "learning_rate": 9.5e-06,
37
+ "loss": 0.4739,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.29,
42
+ "learning_rate": 8.875e-06,
43
+ "loss": 0.4768,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.34,
48
+ "learning_rate": 8.25e-06,
49
+ "loss": 0.4344,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.39,
54
+ "learning_rate": 7.625e-06,
55
+ "loss": 0.4655,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.43,
60
+ "learning_rate": 7e-06,
61
+ "loss": 0.4108,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.48,
66
+ "learning_rate": 6.375e-06,
67
+ "loss": 0.3908,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.48,
72
+ "eval_loss": 0.5148195028305054,
73
+ "eval_runtime": 2783.9171,
74
+ "eval_samples_per_second": 1.186,
75
+ "eval_steps_per_second": 0.074,
76
+ "eval_wer": 30.706069173864208,
77
+ "step": 100
78
+ },
79
+ {
80
+ "epoch": 0.53,
81
+ "learning_rate": 5.75e-06,
82
+ "loss": 0.3863,
83
+ "step": 110
84
+ },
85
+ {
86
+ "epoch": 0.58,
87
+ "learning_rate": 5.125e-06,
88
+ "loss": 0.3902,
89
+ "step": 120
90
+ },
91
+ {
92
+ "epoch": 0.63,
93
+ "learning_rate": 4.5e-06,
94
+ "loss": 0.3788,
95
+ "step": 130
96
+ },
97
+ {
98
+ "epoch": 0.68,
99
+ "learning_rate": 3.875e-06,
100
+ "loss": 0.403,
101
+ "step": 140
102
+ },
103
+ {
104
+ "epoch": 0.72,
105
+ "learning_rate": 3.2500000000000002e-06,
106
+ "loss": 0.3781,
107
+ "step": 150
108
+ },
109
+ {
110
+ "epoch": 0.77,
111
+ "learning_rate": 2.6250000000000003e-06,
112
+ "loss": 0.4094,
113
+ "step": 160
114
+ },
115
+ {
116
+ "epoch": 0.82,
117
+ "learning_rate": 2.0000000000000003e-06,
118
+ "loss": 0.3732,
119
+ "step": 170
120
+ },
121
+ {
122
+ "epoch": 0.87,
123
+ "learning_rate": 1.3750000000000002e-06,
124
+ "loss": 0.4084,
125
+ "step": 180
126
+ },
127
+ {
128
+ "epoch": 0.92,
129
+ "learning_rate": 7.5e-07,
130
+ "loss": 0.3786,
131
+ "step": 190
132
+ },
133
+ {
134
+ "epoch": 0.97,
135
+ "learning_rate": 1.2500000000000002e-07,
136
+ "loss": 0.3519,
137
+ "step": 200
138
+ },
139
+ {
140
+ "epoch": 0.97,
141
+ "eval_loss": 0.46850404143333435,
142
+ "eval_runtime": 2556.1826,
143
+ "eval_samples_per_second": 1.292,
144
+ "eval_steps_per_second": 0.081,
145
+ "eval_wer": 26.980130911344357,
146
+ "step": 200
147
+ },
148
+ {
149
+ "epoch": 0.97,
150
+ "step": 200,
151
+ "total_flos": 6.531871408128e+18,
152
+ "train_loss": 0.4665303599834442,
153
+ "train_runtime": 6285.9399,
154
+ "train_samples_per_second": 1.018,
155
+ "train_steps_per_second": 0.032
156
+ }
157
+ ],
158
+ "max_steps": 200,
159
+ "num_train_epochs": 1,
160
+ "total_flos": 6.531871408128e+18,
161
+ "trial_name": null,
162
+ "trial_params": null
163
+ }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b747d3d49e8fc76cb72a7ae75a9c9a2591850e63223d0fdc354d5c4826934c9f
3
  size 3579
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6406a19600ef55b665d6e036bbd71317a735105e08ff51d4b60484c5c5f0047a
3
  size 3579