End of training

README.md

@@ -1,6 +1,6 @@
 ---
 license: apache-2.0
-base_model: facebook/hubert-base-ls960
+base_model: harshit345/xlsr-wav2vec-speech-emotion-recognition
 tags:
 - generated_from_trainer
 metrics:
@@ -15,10 +15,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ckpts
 
-This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset.
+This model is a fine-tuned version of [harshit345/xlsr-wav2vec-speech-emotion-recognition](https://huggingface.co/harshit345/xlsr-wav2vec-speech-emotion-recognition) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1910
-- Accuracy: 0.9697
+- Loss: 0.2342
+- Accuracy: 0.9333
 
 ## Model description
 
@@ -52,8 +52,8 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss | Accuracy |
 |:-------------:|:-----:|:----:|:---------------:|:--------:|
-| 0.0228        | 2.02  | 500  | 0.1773          | 0.9737   |
-| 0.0385        | 4.04  | 1000 | 0.1910          | 0.9697   |
+| 0.4039        | 2.02  | 500  | 0.3198          | 0.8949   |
+| 0.2195        | 4.04  | 1000 | 0.2363          | 0.9333   |
 
 
 ### Framework versions

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 4.99,
-    "total_flos": 3.160649695892544e+17,
-    "train_loss": 0.03992950008948322,
-    "train_runtime": 666.1547,
+    "total_flos": 1.053348926143392e+18,
+    "train_loss": 0.5674189640925481,
+    "train_runtime": 1664.3033,
     "train_samples": 1979,
-    "train_samples_per_second": 14.854,
-    "train_steps_per_second": 1.854
+    "train_samples_per_second": 5.945,
+    "train_steps_per_second": 0.742
 }

config.json

@@ -1,14 +1,20 @@
 {
-  "_name_or_path": "facebook/hubert-base-ls960",
-  "activation_dropout": 0.1,
+  "_name_or_path": "harshit345/xlsr-wav2vec-speech-emotion-recognition",
+  "activation_dropout": 0.0,
+  "adapter_attn_dim": null,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
   "apply_spec_augment": true,
   "architectures": [
-    "HubertForSpeechClassification"
+    "Wav2Vec2ForSpeechClassification"
   ],
   "attention_dropout": 0.1,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "conv_bias": false,
+  "codevector_dim": 256,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -36,22 +42,21 @@
     2,
     2
   ],
-  "ctc_loss_reduction": "sum",
-  "ctc_zero_infinity": false,
-  "do_stable_layer_norm": false,
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": true,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "group",
-  "feat_proj_dropout": 0.1,
-  "feat_proj_layer_norm": true,
-  "final_dropout": 0.1,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.0,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
   "finetuning_task": "wav2vec2_clf",
-  "gradient_checkpointing": false,
   "hidden_act": "gelu",
   "hidden_dropout": 0.1,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "id2label": {
     "0": "01Neutral",
     "1": "02Fearful",
@@ -60,7 +65,7 @@
     "4": "05Angry"
   },
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "01Neutral": 0,
     "02Fearful": 1,
@@ -70,24 +75,59 @@
   },
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.1,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
   "mask_feature_length": 10,
   "mask_feature_min_masks": 0,
   "mask_feature_prob": 0.0,
   "mask_time_length": 10,
   "mask_time_min_masks": 2,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
   "mask_time_prob": 0.05,
-  "model_type": "hubert",
-  "num_attention_heads": 12,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
   "num_conv_pos_embedding_groups": 16,
   "num_conv_pos_embeddings": 128,
   "num_feat_extract_layers": 7,
-  "num_hidden_layers": 12,
-  "pad_token_id": 0,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
+  "pad_token_id": 54,
   "pooling_mode": "mean",
   "problem_type": "single_label_classification",
-  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "proj_codevector_dim": 256,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
   "torch_dtype": "float32",
   "transformers_version": "4.40.0.dev0",
   "use_weighted_layer_sum": false,
-  "vocab_size": 32
+  "vocab_size": 55,
+  "xvector_output_dim": 512
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d0debed0760df769d15ac8682084547cc2c5e3cc04ab2f3880eafb91150c315
-size 379890236
+oid sha256:4c211db44690f250de2ef2b76fda76600205ce3558af66c9bc6dea154e1a0342
+size 1266026604

preprocessor_config.json

@@ -3,7 +3,7 @@
   "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,
   "padding_side": "right",
-  "padding_value": 0,
-  "return_attention_mask": false,
+  "padding_value": 0.0,
+  "return_attention_mask": true,
   "sampling_rate": 16000
 }

runs/Apr13_05-04-22_af682c374eb8/events.out.tfevents.1712984663.af682c374eb8.34.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c650ee9887f4eba092f7b519c1be298def73b23247f9f5434a6fdf301f44c2df
+size 10230

runs/Apr13_05-04-22_af682c374eb8/events.out.tfevents.1712986374.af682c374eb8.34.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61ab4db88d71a980f11939e7100d324f7d520a81907ed4fad01897c42f2f5820
+size 411

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 4.99,
-    "total_flos": 3.160649695892544e+17,
-    "train_loss": 0.03992950008948322,
-    "train_runtime": 666.1547,
+    "total_flos": 1.053348926143392e+18,
+    "train_loss": 0.5674189640925481,
+    "train_runtime": 1664.3033,
     "train_samples": 1979,
-    "train_samples_per_second": 14.854,
-    "train_steps_per_second": 1.854
+    "train_samples_per_second": 5.945,
+    "train_steps_per_second": 0.742
 }

trainer_state.json

@@ -10,114 +10,114 @@
   "log_history": [
     {
       "epoch": 0.4,
-      "grad_norm": 0.006182590499520302,
+      "grad_norm": 4.1816935539245605,
       "learning_rate": 9.206477732793523e-06,
-      "loss": 0.0729,
+      "loss": 2.284,
       "step": 100
     },
     {
       "epoch": 0.81,
-      "grad_norm": 21.23886489868164,
+      "grad_norm": 7.234564304351807,
       "learning_rate": 8.39676113360324e-06,
-      "loss": 0.0698,
+      "loss": 1.1378,
       "step": 200
     },
     {
       "epoch": 1.21,
-      "grad_norm": 0.006576939485967159,
+      "grad_norm": 4.295170783996582,
       "learning_rate": 7.587044534412956e-06,
-      "loss": 0.0464,
+      "loss": 0.7549,
       "step": 300
     },
     {
       "epoch": 1.62,
-      "grad_norm": 0.009822617284953594,
-      "learning_rate": 6.785425101214575e-06,
-      "loss": 0.0689,
+      "grad_norm": 2.47021746635437,
+      "learning_rate": 6.777327935222673e-06,
+      "loss": 0.545,
       "step": 400
     },
     {
       "epoch": 2.02,
-      "grad_norm": 0.00560363894328475,
-      "learning_rate": 5.975708502024292e-06,
-      "loss": 0.0228,
+      "grad_norm": 4.9068498611450195,
+      "learning_rate": 5.9676113360323896e-06,
+      "loss": 0.4039,
       "step": 500
     },
     {
       "epoch": 2.02,
-      "eval_accuracy": 0.973737359046936,
-      "eval_loss": 0.17728736996650696,
-      "eval_runtime": 23.1344,
-      "eval_samples_per_second": 21.397,
-      "eval_steps_per_second": 5.36,
+      "eval_accuracy": 0.8949494957923889,
+      "eval_loss": 0.3197959065437317,
+      "eval_runtime": 43.5773,
+      "eval_samples_per_second": 11.359,
+      "eval_steps_per_second": 2.846,
       "step": 500
     },
     {
       "epoch": 2.42,
-      "grad_norm": 0.010259171947836876,
-      "learning_rate": 5.165991902834009e-06,
-      "loss": 0.0253,
+      "grad_norm": 10.178821563720703,
+      "learning_rate": 5.157894736842106e-06,
+      "loss": 0.3572,
       "step": 600
     },
     {
       "epoch": 2.83,
-      "grad_norm": 0.5489621758460999,
-      "learning_rate": 4.356275303643725e-06,
-      "loss": 0.0227,
+      "grad_norm": 22.03152847290039,
+      "learning_rate": 4.348178137651822e-06,
+      "loss": 0.3237,
       "step": 700
     },
     {
       "epoch": 3.23,
-      "grad_norm": 0.003071287414059043,
-      "learning_rate": 3.5465587044534415e-06,
-      "loss": 0.0398,
+      "grad_norm": 4.856521129608154,
+      "learning_rate": 3.538461538461539e-06,
+      "loss": 0.2635,
       "step": 800
     },
     {
       "epoch": 3.64,
-      "grad_norm": 0.0034055381547659636,
+      "grad_norm": 4.2341766357421875,
       "learning_rate": 2.7368421052631583e-06,
-      "loss": 0.0299,
+      "loss": 0.2138,
       "step": 900
     },
     {
       "epoch": 4.04,
-      "grad_norm": 0.005517472513020039,
+      "grad_norm": 0.4156922996044159,
       "learning_rate": 1.9271255060728746e-06,
-      "loss": 0.0385,
+      "loss": 0.2195,
       "step": 1000
     },
     {
       "epoch": 4.04,
-      "eval_accuracy": 0.9696969985961914,
-      "eval_loss": 0.190963014960289,
-      "eval_runtime": 23.4329,
-      "eval_samples_per_second": 21.124,
-      "eval_steps_per_second": 5.292,
+      "eval_accuracy": 0.9333333373069763,
+      "eval_loss": 0.2362980842590332,
+      "eval_runtime": 43.7442,
+      "eval_samples_per_second": 11.316,
+      "eval_steps_per_second": 2.835,
       "step": 1000
     },
     {
       "epoch": 4.44,
-      "grad_norm": 19.359519958496094,
+      "grad_norm": 5.4564337730407715,
       "learning_rate": 1.1174089068825912e-06,
-      "loss": 0.0365,
+      "loss": 0.2312,
       "step": 1100
     },
     {
       "epoch": 4.85,
-      "grad_norm": 1.323317527770996,
+      "grad_norm": 2.0700230598449707,
       "learning_rate": 3.0769230769230774e-07,
-      "loss": 0.0064,
+      "loss": 0.182,
       "step": 1200
     },
     {
       "epoch": 4.99,
       "step": 1235,
-      "total_flos": 3.160649695892544e+17,
-      "train_loss": 0.03992950008948322,
-      "train_runtime": 666.1547,
-      "train_samples_per_second": 14.854,
-      "train_steps_per_second": 1.854
+      "total_flos": 1.053348926143392e+18,
+      "train_loss": 0.5674189640925481,
+      "train_runtime": 1664.3033,
+      "train_samples_per_second": 5.945,
+      "train_steps_per_second": 0.742
     }
   ],
   "logging_steps": 100,
@@ -125,7 +125,7 @@
   "num_input_tokens_seen": 0,
   "num_train_epochs": 5,
   "save_steps": 500,
-  "total_flos": 3.160649695892544e+17,
+  "total_flos": 1.053348926143392e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f507758a89624d13b7bed17a703fed0e8a853809c160f6b3fee91839d462fc37
+oid sha256:6aad4c2aa7f6067110d61ed613d6e4dbae9214f7578b02e94a2191162bee5aa5
 size 4984