yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor

README.md

@@ -3,8 +3,6 @@ license: apache-2.0
 base_model: yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor
 tags:
 - generated_from_trainer
-- This model is fully trained on the  sample-Data of 11 languages provided
-- not trained yet TTS-data or the second one
 metrics:
 - accuracy
 model-index:
@@ -19,17 +17,16 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor](https://huggingface.co/yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1719
-- Accuracy: 0.9554
+- Loss: 0.4481
+- Accuracy: 0.8710
 
 ## Model description
 
-This is final finetuned version of Wav2vec2 on the sample data of audio from 11 Indian Languages codemixed with English language.
-It has only been yet finetuned on one dataset that is sample dataset.
+More information needed
 
 ## Intended uses & limitations
 
-Not giving yet good accuracy on other datasets like TTS and sample dataset. Needs to be trained on more dataset.
+More information needed
 
 ## Training and evaluation data
 
@@ -41,28 +38,22 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 5e-05
-- train_batch_size: 16
-- eval_batch_size: 16
+- train_batch_size: 8
+- eval_batch_size: 8
 - seed: 42
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 256
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 64
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- num_epochs: 90
+- num_epochs: 60
 
 ### Training results
 
-| Training Loss | Epoch | Step | Validation Loss | Accuracy |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 0.0103        | 21.11 | 1000 | 0.1802          | 0.9501   |
-| 0.009         | 42.22 | 2000 | 0.1717          | 0.9497   |
-| 0.0086        | 63.32 | 3000 | 0.1675          | 0.9546   |
-| 0.0073        | 84.43 | 4000 | 0.1686          | 0.9538   |
 
 
 ### Framework versions
 
-- Transformers 4.33.0
+- Transformers 4.32.1
 - Pytorch 2.0.0
 - Datasets 2.11.0
 - Tokenizers 0.13.3

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 89.29,
-    "eval_accuracy": 0.9554455280303955,
-    "eval_loss": 0.1719195693731308,
-    "eval_runtime": 50.8715,
-    "eval_samples": 2424,
-    "eval_samples_per_second": 47.65,
-    "eval_steps_per_second": 2.988,
-    "total_flos": 3.2880550437308154e+19,
-    "train_loss": 0.00991302564845863,
-    "train_runtime": 33902.9866,
-    "train_samples": 12120,
-    "train_samples_per_second": 32.174,
-    "train_steps_per_second": 0.125
+    "epoch": 56.95,
+    "eval_accuracy": 0.8709677457809448,
+    "eval_loss": 0.4480999708175659,
+    "eval_runtime": 1.4774,
+    "eval_samples": 93,
+    "eval_samples_per_second": 62.95,
+    "eval_steps_per_second": 8.123,
+    "total_flos": 8.064772262536032e+17,
+    "train_loss": 0.15708597316628412,
+    "train_runtime": 790.3622,
+    "train_samples": 466,
+    "train_samples_per_second": 35.376,
+    "train_steps_per_second": 0.531
 }

config.json

@@ -139,7 +139,7 @@
     1
   ],
   "torch_dtype": "float32",
-  "transformers_version": "4.33.0",
+  "transformers_version": "4.32.1",
   "use_weighted_layer_sum": false,
   "vocab_size": 32,
   "xvector_output_dim": 512

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 89.29,
-    "eval_accuracy": 0.9554455280303955,
-    "eval_loss": 0.1719195693731308,
-    "eval_runtime": 50.8715,
-    "eval_samples": 2424,
-    "eval_samples_per_second": 47.65,
-    "eval_steps_per_second": 2.988
+    "epoch": 56.95,
+    "eval_accuracy": 0.8709677457809448,
+    "eval_loss": 0.4480999708175659,
+    "eval_runtime": 1.4774,
+    "eval_samples": 93,
+    "eval_samples_per_second": 62.95,
+    "eval_steps_per_second": 8.123
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7ceaacefa2250a52d0d79eeb85b6a0da21680d9b9b79e1d64c35a9ab6bd911c1
+oid sha256:c5bd079f25010ad6fed494f319bd825fbda1ad3ba0247b15ee3faa0fc2a04cef
 size 1266146037

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 89.29,
-    "total_flos": 3.2880550437308154e+19,
-    "train_loss": 0.00991302564845863,
-    "train_runtime": 33902.9866,
-    "train_samples": 12120,
-    "train_samples_per_second": 32.174,
-    "train_steps_per_second": 0.125
+    "epoch": 56.95,
+    "total_flos": 8.064772262536032e+17,
+    "train_loss": 0.15708597316628412,
+    "train_runtime": 790.3622,
+    "train_samples": 466,
+    "train_samples_per_second": 35.376,
+    "train_steps_per_second": 0.531
 }

trainer_state.json

@@ -1,325 +1,61 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 89.28759894459102,
+  "epoch": 56.94915254237288,
   "eval_steps": 1000,
-  "global_step": 4230,
+  "global_step": 420,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 2.11,
-      "learning_rate": 4.8817966903073283e-05,
-      "loss": 0.0118,
+      "epoch": 13.56,
+      "learning_rate": 3.821428571428572e-05,
+      "loss": 0.554,
       "step": 100
     },
     {
-      "epoch": 4.22,
-      "learning_rate": 4.763593380614658e-05,
-      "loss": 0.0156,
+      "epoch": 27.12,
+      "learning_rate": 2.6309523809523813e-05,
+      "loss": 0.0396,
       "step": 200
     },
     {
-      "epoch": 6.33,
-      "learning_rate": 4.645390070921986e-05,
-      "loss": 0.0122,
+      "epoch": 40.68,
+      "learning_rate": 1.4404761904761905e-05,
+      "loss": 0.0312,
       "step": 300
     },
     {
-      "epoch": 8.44,
-      "learning_rate": 4.527186761229315e-05,
-      "loss": 0.0148,
+      "epoch": 54.24,
+      "learning_rate": 2.5e-06,
+      "loss": 0.0308,
       "step": 400
     },
     {
-      "epoch": 10.55,
-      "learning_rate": 4.4089834515366435e-05,
-      "loss": 0.0114,
-      "step": 500
+      "epoch": 56.95,
+      "step": 420,
+      "total_flos": 8.064772262536032e+17,
+      "train_loss": 0.15708597316628412,
+      "train_runtime": 790.3622,
+      "train_samples_per_second": 35.376,
+      "train_steps_per_second": 0.531
     },
     {
-      "epoch": 12.66,
-      "learning_rate": 4.2907801418439716e-05,
-      "loss": 0.0143,
-      "step": 600
-    },
-    {
-      "epoch": 14.78,
-      "learning_rate": 4.1725768321513004e-05,
-      "loss": 0.0149,
-      "step": 700
-    },
-    {
-      "epoch": 16.89,
-      "learning_rate": 4.0543735224586285e-05,
-      "loss": 0.0139,
-      "step": 800
-    },
-    {
-      "epoch": 19.0,
-      "learning_rate": 3.936170212765958e-05,
-      "loss": 0.0125,
-      "step": 900
-    },
-    {
-      "epoch": 21.11,
-      "learning_rate": 3.817966903073286e-05,
-      "loss": 0.0103,
-      "step": 1000
-    },
-    {
-      "epoch": 21.11,
-      "eval_accuracy": 0.9500824809074402,
-      "eval_loss": 0.18024244904518127,
-      "eval_runtime": 47.0038,
-      "eval_samples_per_second": 51.57,
-      "eval_steps_per_second": 3.234,
-      "step": 1000
-    },
-    {
-      "epoch": 23.22,
-      "learning_rate": 3.699763593380615e-05,
-      "loss": 0.0111,
-      "step": 1100
-    },
-    {
-      "epoch": 25.33,
-      "learning_rate": 3.5815602836879437e-05,
-      "loss": 0.0093,
-      "step": 1200
-    },
-    {
-      "epoch": 27.44,
-      "learning_rate": 3.463356973995272e-05,
-      "loss": 0.0109,
-      "step": 1300
-    },
-    {
-      "epoch": 29.55,
-      "learning_rate": 3.3451536643026005e-05,
-      "loss": 0.0102,
-      "step": 1400
-    },
-    {
-      "epoch": 31.66,
-      "learning_rate": 3.226950354609929e-05,
-      "loss": 0.012,
-      "step": 1500
-    },
-    {
-      "epoch": 33.77,
-      "learning_rate": 3.108747044917258e-05,
-      "loss": 0.0116,
-      "step": 1600
-    },
-    {
-      "epoch": 35.88,
-      "learning_rate": 2.9905437352245862e-05,
-      "loss": 0.0145,
-      "step": 1700
-    },
-    {
-      "epoch": 37.99,
-      "learning_rate": 2.8723404255319154e-05,
-      "loss": 0.011,
-      "step": 1800
-    },
-    {
-      "epoch": 40.11,
-      "learning_rate": 2.7541371158392438e-05,
-      "loss": 0.0108,
-      "step": 1900
-    },
-    {
-      "epoch": 42.22,
-      "learning_rate": 2.6359338061465723e-05,
-      "loss": 0.009,
-      "step": 2000
-    },
-    {
-      "epoch": 42.22,
-      "eval_accuracy": 0.9496699571609497,
-      "eval_loss": 0.1716560274362564,
-      "eval_runtime": 51.12,
-      "eval_samples_per_second": 47.418,
-      "eval_steps_per_second": 2.973,
-      "step": 2000
-    },
-    {
-      "epoch": 44.33,
-      "learning_rate": 2.5177304964539007e-05,
-      "loss": 0.0101,
-      "step": 2100
-    },
-    {
-      "epoch": 46.44,
-      "learning_rate": 2.3995271867612295e-05,
-      "loss": 0.0087,
-      "step": 2200
-    },
-    {
-      "epoch": 48.55,
-      "learning_rate": 2.281323877068558e-05,
-      "loss": 0.0114,
-      "step": 2300
-    },
-    {
-      "epoch": 50.66,
-      "learning_rate": 2.1631205673758867e-05,
-      "loss": 0.0076,
-      "step": 2400
-    },
-    {
-      "epoch": 52.77,
-      "learning_rate": 2.0449172576832152e-05,
-      "loss": 0.0088,
-      "step": 2500
-    },
-    {
-      "epoch": 54.88,
-      "learning_rate": 1.926713947990544e-05,
-      "loss": 0.0084,
-      "step": 2600
-    },
-    {
-      "epoch": 56.99,
-      "learning_rate": 1.8085106382978724e-05,
-      "loss": 0.0095,
-      "step": 2700
-    },
-    {
-      "epoch": 59.1,
-      "learning_rate": 1.690307328605201e-05,
-      "loss": 0.0075,
-      "step": 2800
-    },
-    {
-      "epoch": 61.21,
-      "learning_rate": 1.5721040189125296e-05,
-      "loss": 0.0097,
-      "step": 2900
-    },
-    {
-      "epoch": 63.32,
-      "learning_rate": 1.4539007092198581e-05,
-      "loss": 0.0086,
-      "step": 3000
-    },
-    {
-      "epoch": 63.32,
-      "eval_accuracy": 0.9546204805374146,
-      "eval_loss": 0.16754871606826782,
-      "eval_runtime": 52.1005,
-      "eval_samples_per_second": 46.525,
-      "eval_steps_per_second": 2.917,
-      "step": 3000
-    },
-    {
-      "epoch": 65.44,
-      "learning_rate": 1.3356973995271869e-05,
-      "loss": 0.0079,
-      "step": 3100
-    },
-    {
-      "epoch": 67.55,
-      "learning_rate": 1.2174940898345153e-05,
-      "loss": 0.0076,
-      "step": 3200
-    },
-    {
-      "epoch": 69.66,
-      "learning_rate": 1.0992907801418441e-05,
-      "loss": 0.0072,
-      "step": 3300
-    },
-    {
-      "epoch": 71.77,
-      "learning_rate": 9.810874704491727e-06,
-      "loss": 0.0074,
-      "step": 3400
-    },
-    {
-      "epoch": 73.88,
-      "learning_rate": 8.628841607565012e-06,
-      "loss": 0.0076,
-      "step": 3500
-    },
-    {
-      "epoch": 75.99,
-      "learning_rate": 7.446808510638298e-06,
-      "loss": 0.0069,
-      "step": 3600
-    },
-    {
-      "epoch": 78.1,
-      "learning_rate": 6.264775413711583e-06,
-      "loss": 0.0068,
-      "step": 3700
-    },
-    {
-      "epoch": 80.21,
-      "learning_rate": 5.08274231678487e-06,
-      "loss": 0.007,
-      "step": 3800
-    },
-    {
-      "epoch": 82.32,
-      "learning_rate": 3.9007092198581565e-06,
-      "loss": 0.0072,
-      "step": 3900
-    },
-    {
-      "epoch": 84.43,
-      "learning_rate": 2.7186761229314422e-06,
-      "loss": 0.0073,
-      "step": 4000
-    },
-    {
-      "epoch": 84.43,
-      "eval_accuracy": 0.9537953734397888,
-      "eval_loss": 0.16863200068473816,
-      "eval_runtime": 51.419,
-      "eval_samples_per_second": 47.142,
-      "eval_steps_per_second": 2.956,
-      "step": 4000
-    },
-    {
-      "epoch": 86.54,
-      "learning_rate": 1.5484633569739953e-06,
-      "loss": 0.0065,
-      "step": 4100
-    },
-    {
-      "epoch": 88.65,
-      "learning_rate": 3.6643026004728135e-07,
-      "loss": 0.0058,
-      "step": 4200
-    },
-    {
-      "epoch": 89.29,
-      "step": 4230,
-      "total_flos": 3.2880550437308154e+19,
-      "train_loss": 0.00991302564845863,
-      "train_runtime": 33902.9866,
-      "train_samples_per_second": 32.174,
-      "train_steps_per_second": 0.125
-    },
-    {
-      "epoch": 89.29,
-      "eval_accuracy": 0.9554455280303955,
-      "eval_loss": 0.1719195693731308,
-      "eval_runtime": 50.8715,
-      "eval_samples_per_second": 47.65,
-      "eval_steps_per_second": 2.988,
-      "step": 4230
+      "epoch": 56.95,
+      "eval_accuracy": 0.8709677457809448,
+      "eval_loss": 0.4480999708175659,
+      "eval_runtime": 1.4774,
+      "eval_samples_per_second": 62.95,
+      "eval_steps_per_second": 8.123,
+      "step": 420
     }
   ],
   "logging_steps": 100,
-  "max_steps": 4230,
-  "num_train_epochs": 90,
+  "max_steps": 420,
+  "num_train_epochs": 60,
   "save_steps": 2000,
-  "total_flos": 3.2880550437308154e+19,
+  "total_flos": 8.064772262536032e+17,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afafbb8d751fc33fbf51d1298497b2fbbc858aa6e7af5c8ee9fc1310c74fcc53
+oid sha256:a6fff3406fc6d17e7151844526156d27d071a854fa3b738d788067583d864923
 size 4155