Upload folder using huggingface_hub

README.md

@@ -0,0 +1,86 @@
+---
+license: apache-2.0
+base_model: facebook/hubert-xlarge-ll60k
+tags:
+- audio-classification
+- generated_from_trainer
+datasets:
+- superb
+metrics:
+- accuracy
+model-index:
+- name: superb_ks_42
+  results:
+  - task:
+      name: Audio Classification
+      type: audio-classification
+    dataset:
+      name: superb
+      type: superb
+      config: ks
+      split: validation
+      args: ks
+    metrics:
+    - name: Accuracy
+      type: accuracy
+      value: 0.9872021182700794
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# superb_ks_42
+
+This model is a fine-tuned version of [facebook/hubert-xlarge-ll60k](https://huggingface.co/facebook/hubert-xlarge-ll60k) on the superb dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0579
+- Accuracy: 0.9872
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 5e-05
+- train_batch_size: 32
+- eval_batch_size: 4
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 10
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
+|:-------------:|:-----:|:-----:|:---------------:|:--------:|
+| 1.5639        | 1.0   | 1597  | 0.1012          | 0.9794   |
+| 0.205         | 2.0   | 3194  | 0.0720          | 0.9841   |
+| 0.1774        | 3.0   | 4791  | 0.0664          | 0.9856   |
+| 0.1549        | 4.0   | 6388  | 0.0621          | 0.9856   |
+| 0.1416        | 5.0   | 7985  | 0.0620          | 0.9859   |
+| 0.1328        | 6.0   | 9582  | 0.0614          | 0.9865   |
+| 0.1159        | 7.0   | 11179 | 0.0615          | 0.9865   |
+| 0.1211        | 8.0   | 12776 | 0.0579          | 0.9872   |
+| 0.108         | 9.0   | 14373 | 0.0566          | 0.9863   |
+| 0.1088        | 10.0  | 15970 | 0.0572          | 0.9869   |
+
+
+### Framework versions
+
+- Transformers 4.40.1
+- Pytorch 2.3.0+cu121
+- Datasets 2.19.0
+- Tokenizers 0.19.1

all_results.json

@@ -0,0 +1,19 @@
+{
+    "epoch": 10.0,
+    "eval_accuracy": 0.9872021182700794,
+    "eval_loss": 0.057863999158144,
+    "eval_runtime": 34.232,
+    "eval_samples_per_second": 198.586,
+    "eval_steps_per_second": 0.789,
+    "test_accuracy": 0.8980850373255437,
+    "test_loss": 0.5234816074371338,
+    "test_runtime": 16.7699,
+    "test_samples_per_second": 183.722,
+    "test_steps_per_second": 0.775,
+    "total_flos": 4.722696547222273e+19,
+    "train_accuracy": 0.9925040122127843,
+    "train_loss": 0.03099818341434002,
+    "train_runtime": 461.453,
+    "train_samples_per_second": 110.724,
+    "train_steps_per_second": 27.682
+}

config.json

@@ -0,0 +1,112 @@
+{
+  "_name_or_path": "facebook/hubert-xlarge-ll60k",
+  "activation_dropout": 0.0,
+  "apply_spec_augment": true,
+  "architectures": [
+    "HubertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "classifier_proj_size": 256,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "sum",
+  "ctc_zero_infinity": false,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.1,
+  "feat_proj_layer_norm": true,
+  "final_dropout": 0.0,
+  "finetuning_task": "audio-classification",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 1280,
+  "id2label": {
+    "0": "yes",
+    "1": "no",
+    "10": "_silence_",
+    "11": "_unknown_",
+    "2": "up",
+    "3": "down",
+    "4": "left",
+    "5": "right",
+    "6": "on",
+    "7": "off",
+    "8": "stop",
+    "9": "go"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 5120,
+  "label2id": {
+    "_silence_": "10",
+    "_unknown_": "11",
+    "down": "3",
+    "go": "9",
+    "left": "4",
+    "no": "1",
+    "off": "7",
+    "on": "6",
+    "right": "5",
+    "stop": "8",
+    "up": "2",
+    "yes": "0"
+  },
+  "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": 0.075,
+  "mask_time_selection": "static",
+  "model_type": "hubert",
+  "num_attention_heads": 16,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 48,
+  "pad_token_id": 0,
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.40.1",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 32
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 10.0,
+    "eval_accuracy": 0.9872021182700794,
+    "eval_loss": 0.057863999158144,
+    "eval_runtime": 34.232,
+    "eval_samples_per_second": 198.586,
+    "eval_steps_per_second": 0.789
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f2bc11bb0ef6d16449d224dc914cc834d3e8c73c64c915023349c2c379afe6b
+size 3851413640

pred_logit_eval.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:515d91afcfdd4ffaccd8dff308a92766c5f64db77ed38f90b4ba17fe9f66ae09
+size 326432

pred_logit_test.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d3062802296c6afefef5c4e321e7a68cd3be760c2c228def7dfa921b6b9cd1d
+size 148016

pred_logit_train.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3adaa2dea82eb143bea7578da601c7303fc71183d7b1dc5143e2a8285263064
+size 2452640

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

test_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 10.0,
+    "test_accuracy": 0.8980850373255437,
+    "test_loss": 0.5234816074371338,
+    "test_runtime": 16.7699,
+    "test_samples_per_second": 183.722,
+    "test_steps_per_second": 0.775
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 10.0,
+    "train_accuracy": 0.9925040122127843,
+    "train_loss": 0.03099818341434002,
+    "train_runtime": 461.453,
+    "train_samples_per_second": 110.724,
+    "train_steps_per_second": 27.682
+}

trainer_state.json

@@ -0,0 +1,225 @@
+{
+  "best_metric": 0.9872021182700794,
+  "best_model_checkpoint": "audio/train/checkpoint/hubert-xlarge/superb_ks_42/checkpoint-12776",
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 15970,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.6261740763932373,
+      "grad_norm": 4.316092491149902,
+      "learning_rate": 3.1308703819661866e-05,
+      "loss": 1.5639,
+      "step": 1000
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.9794057075610474,
+      "eval_loss": 0.10121691972017288,
+      "eval_runtime": 36.0525,
+      "eval_samples_per_second": 188.558,
+      "eval_steps_per_second": 0.749,
+      "step": 1597
+    },
+    {
+      "epoch": 1.2523481527864746,
+      "grad_norm": 1.4483940601348877,
+      "learning_rate": 4.859806581785292e-05,
+      "loss": 0.3591,
+      "step": 2000
+    },
+    {
+      "epoch": 1.878522229179712,
+      "grad_norm": 0.9219438433647156,
+      "learning_rate": 4.51193209490016e-05,
+      "loss": 0.205,
+      "step": 3000
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9841129744042365,
+      "eval_loss": 0.07202675193548203,
+      "eval_runtime": 35.6833,
+      "eval_samples_per_second": 190.509,
+      "eval_steps_per_second": 0.757,
+      "step": 3194
+    },
+    {
+      "epoch": 2.504696305572949,
+      "grad_norm": 1.6926661729812622,
+      "learning_rate": 4.1640576080150286e-05,
+      "loss": 0.1774,
+      "step": 4000
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.9855839952927331,
+      "eval_loss": 0.06639665365219116,
+      "eval_runtime": 35.3875,
+      "eval_samples_per_second": 192.102,
+      "eval_steps_per_second": 0.763,
+      "step": 4791
+    },
+    {
+      "epoch": 3.1308703819661865,
+      "grad_norm": 1.2424182891845703,
+      "learning_rate": 3.8161831211298965e-05,
+      "loss": 0.159,
+      "step": 5000
+    },
+    {
+      "epoch": 3.757044458359424,
+      "grad_norm": 0.05382406339049339,
+      "learning_rate": 3.468308634244764e-05,
+      "loss": 0.1549,
+      "step": 6000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9855839952927331,
+      "eval_loss": 0.06212027370929718,
+      "eval_runtime": 35.7277,
+      "eval_samples_per_second": 190.273,
+      "eval_steps_per_second": 0.756,
+      "step": 6388
+    },
+    {
+      "epoch": 4.383218534752661,
+      "grad_norm": 0.4753202199935913,
+      "learning_rate": 3.120434147359633e-05,
+      "loss": 0.1416,
+      "step": 7000
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9858781994704324,
+      "eval_loss": 0.06196163594722748,
+      "eval_runtime": 35.6118,
+      "eval_samples_per_second": 190.892,
+      "eval_steps_per_second": 0.758,
+      "step": 7985
+    },
+    {
+      "epoch": 5.009392611145898,
+      "grad_norm": 1.2926490306854248,
+      "learning_rate": 2.7725596604745007e-05,
+      "loss": 0.1378,
+      "step": 8000
+    },
+    {
+      "epoch": 5.635566687539136,
+      "grad_norm": 0.03198292478919029,
+      "learning_rate": 2.4246851735893692e-05,
+      "loss": 0.1328,
+      "step": 9000
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.9864666078258312,
+      "eval_loss": 0.061353616416454315,
+      "eval_runtime": 35.411,
+      "eval_samples_per_second": 191.974,
+      "eval_steps_per_second": 0.762,
+      "step": 9582
+    },
+    {
+      "epoch": 6.261740763932373,
+      "grad_norm": 1.4529321193695068,
+      "learning_rate": 2.076810686704237e-05,
+      "loss": 0.1244,
+      "step": 10000
+    },
+    {
+      "epoch": 6.88791484032561,
+      "grad_norm": 1.9151487350463867,
+      "learning_rate": 1.7289361998191053e-05,
+      "loss": 0.1159,
+      "step": 11000
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.9864666078258312,
+      "eval_loss": 0.06154850497841835,
+      "eval_runtime": 35.249,
+      "eval_samples_per_second": 192.857,
+      "eval_steps_per_second": 0.766,
+      "step": 11179
+    },
+    {
+      "epoch": 7.514088916718848,
+      "grad_norm": 4.778970241546631,
+      "learning_rate": 1.3810617129339735e-05,
+      "loss": 0.1211,
+      "step": 12000
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.9872021182700794,
+      "eval_loss": 0.057863999158144,
+      "eval_runtime": 35.7046,
+      "eval_samples_per_second": 190.396,
+      "eval_steps_per_second": 0.756,
+      "step": 12776
+    },
+    {
+      "epoch": 8.140262993112085,
+      "grad_norm": 0.4232797622680664,
+      "learning_rate": 1.0331872260488417e-05,
+      "loss": 0.114,
+      "step": 13000
+    },
+    {
+      "epoch": 8.766437069505322,
+      "grad_norm": 1.9394092559814453,
+      "learning_rate": 6.853127391637097e-06,
+      "loss": 0.108,
+      "step": 14000
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.9863195057369815,
+      "eval_loss": 0.05663929507136345,
+      "eval_runtime": 35.5366,
+      "eval_samples_per_second": 191.296,
+      "eval_steps_per_second": 0.76,
+      "step": 14373
+    },
+    {
+      "epoch": 9.39261114589856,
+      "grad_norm": 3.1410608291625977,
+      "learning_rate": 3.3743825227857787e-06,
+      "loss": 0.1088,
+      "step": 15000
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.9869079140923801,
+      "eval_loss": 0.05716182291507721,
+      "eval_runtime": 35.7353,
+      "eval_samples_per_second": 190.232,
+      "eval_steps_per_second": 0.756,
+      "step": 15970
+    },
+    {
+      "epoch": 10.0,
+      "step": 15970,
+      "total_flos": 4.722696547222273e+19,
+      "train_loss": 0.23973732257382005,
+      "train_runtime": 8088.8257,
+      "train_samples_per_second": 63.166,
+      "train_steps_per_second": 1.974
+    }
+  ],
+  "logging_steps": 1000,
+  "max_steps": 15970,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "total_flos": 4.722696547222273e+19,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3268bf7263d4d82c66b99ba83c1f660d512b945e42e958b041e7fb840fce18d5
+size 5112