Added model files

README.md

@@ -0,0 +1,89 @@
+---
+language:
+- eo
+license: apache-2.0
+tags:
+- automatic-speech-recognition
+- robust-speech-event
+- common_voice
+- generated_from_trainer
+datasets:
+- common_voice
+model-index:
+- name: wav2vec2-xls-r-300m-eo
+  results: []
+---
+
+<!-- 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. -->
+
+# wav2vec2-xls-r-300m-eo
+
+This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the COMMON_VOICE - EO dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2584
+- Wer: 0.3114
+
+## 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: 0.0003
+- train_batch_size: 16
+- eval_batch_size: 8
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 20.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+| Training Loss | Epoch | Step  | Validation Loss | Wer    |
+|:-------------:|:-----:|:-----:|:---------------:|:------:|
+| 3.1701        | 0.8   | 500   | 2.8105          | 1.0    |
+| 1.9143        | 1.6   | 1000  | 0.5977          | 0.7002 |
+| 1.1259        | 2.4   | 1500  | 0.5063          | 0.6157 |
+| 0.9732        | 3.2   | 2000  | 0.4264          | 0.5673 |
+| 0.8983        | 4.0   | 2500  | 0.4249          | 0.4902 |
+| 0.8507        | 4.8   | 3000  | 0.3811          | 0.4536 |
+| 0.8064        | 5.6   | 3500  | 0.3643          | 0.4467 |
+| 0.7866        | 6.4   | 4000  | 0.3600          | 0.4453 |
+| 0.7773        | 7.2   | 4500  | 0.3724          | 0.4470 |
+| 0.747         | 8.0   | 5000  | 0.3501          | 0.4189 |
+| 0.7279        | 8.8   | 5500  | 0.3500          | 0.4261 |
+| 0.7153        | 9.6   | 6000  | 0.3328          | 0.3966 |
+| 0.7           | 10.4  | 6500  | 0.3314          | 0.3869 |
+| 0.6784        | 11.2  | 7000  | 0.3396          | 0.4051 |
+| 0.6582        | 12.0  | 7500  | 0.3236          | 0.3899 |
+| 0.6478        | 12.8  | 8000  | 0.3263          | 0.3832 |
+| 0.6277        | 13.6  | 8500  | 0.3139          | 0.3769 |
+| 0.6053        | 14.4  | 9000  | 0.2955          | 0.3536 |
+| 0.5777        | 15.2  | 9500  | 0.2793          | 0.3413 |
+| 0.5631        | 16.0  | 10000 | 0.2789          | 0.3353 |
+| 0.5446        | 16.8  | 10500 | 0.2709          | 0.3264 |
+| 0.528         | 17.6  | 11000 | 0.2693          | 0.3234 |
+| 0.5169        | 18.4  | 11500 | 0.2656          | 0.3193 |
+| 0.5041        | 19.2  | 12000 | 0.2575          | 0.3102 |
+| 0.4971        | 20.0  | 12500 | 0.2584          | 0.3114 |
+
+
+### Framework versions
+
+- Transformers 4.17.0.dev0
+- Pytorch 1.10.2+cu102
+- Datasets 1.18.2.dev0
+- Tokenizers 0.11.0

added_tokens.json

@@ -0,0 +1 @@
+{"<s>": 55, "</s>": 56}

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 20.0,
+    "eval_loss": 0.2584246098995209,
+    "eval_runtime": 235.5674,
+    "eval_samples": 5000,
+    "eval_samples_per_second": 21.225,
+    "eval_steps_per_second": 2.653,
+    "eval_wer": 0.31141699083267593,
+    "train_loss": 0.8456691796875,
+    "train_runtime": 20148.7553,
+    "train_samples": 10000,
+    "train_samples_per_second": 9.926,
+    "train_steps_per_second": 0.62
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 20.0,
+    "eval_loss": 0.2584246098995209,
+    "eval_runtime": 235.5674,
+    "eval_samples": 5000,
+    "eval_samples_per_second": 21.225,
+    "eval_steps_per_second": 2.653,
+    "eval_wer": 0.31141699083267593
+}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2bf3aed7169f1b2cc746039e640bd88b4d53c287920faa6743254e924058776d
+oid sha256:9f3326278c8c0c1ab021b5c56c7e23a0a60336147e6585104e036ef2b55b33e7
 size 1262157361

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "./wav2vec2-xls-r-300m-eo", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 20.0,
+    "train_loss": 0.8456691796875,
+    "train_runtime": 20148.7553,
+    "train_samples": 10000,
+    "train_samples_per_second": 9.926,
+    "train_steps_per_second": 0.62
+}

trainer_state.json

@@ -0,0 +1,400 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 20.0,
+  "global_step": 12500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00028807199999999995,
+      "loss": 3.1701,
+      "step": 500
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.8104631900787354,
+      "eval_runtime": 249.0567,
+      "eval_samples_per_second": 20.076,
+      "eval_steps_per_second": 2.509,
+      "eval_wer": 1.0,
+      "step": 500
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.000276072,
+      "loss": 1.9143,
+      "step": 1000
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 0.5976669192314148,
+      "eval_runtime": 247.4077,
+      "eval_samples_per_second": 20.21,
+      "eval_steps_per_second": 2.526,
+      "eval_wer": 0.7001713772748293,
+      "step": 1000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 0.00026407199999999996,
+      "loss": 1.1259,
+      "step": 1500
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 0.5062595009803772,
+      "eval_runtime": 241.7474,
+      "eval_samples_per_second": 20.683,
+      "eval_steps_per_second": 2.585,
+      "eval_wer": 0.6156796605097794,
+      "step": 1500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 0.000252072,
+      "loss": 0.9732,
+      "step": 2000
+    },
+    {
+      "epoch": 3.2,
+      "eval_loss": 0.42641571164131165,
+      "eval_runtime": 234.1368,
+      "eval_samples_per_second": 21.355,
+      "eval_steps_per_second": 2.669,
+      "eval_wer": 0.5673403879111014,
+      "step": 2000
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 0.000240096,
+      "loss": 0.8983,
+      "step": 2500
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.42490196228027344,
+      "eval_runtime": 237.618,
+      "eval_samples_per_second": 21.042,
+      "eval_steps_per_second": 2.63,
+      "eval_wer": 0.49024781697995157,
+      "step": 2500
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 0.00022809599999999998,
+      "loss": 0.8507,
+      "step": 3000
+    },
+    {
+      "epoch": 4.8,
+      "eval_loss": 0.38109204173088074,
+      "eval_runtime": 243.3726,
+      "eval_samples_per_second": 20.545,
+      "eval_steps_per_second": 2.568,
+      "eval_wer": 0.4535785207148881,
+      "step": 3000
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 0.00021609599999999996,
+      "loss": 0.8064,
+      "step": 3500
+    },
+    {
+      "epoch": 5.6,
+      "eval_loss": 0.3643375337123871,
+      "eval_runtime": 241.9721,
+      "eval_samples_per_second": 20.664,
+      "eval_steps_per_second": 2.583,
+      "eval_wer": 0.44669622697967953,
+      "step": 3500
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 0.000204096,
+      "loss": 0.7866,
+      "step": 4000
+    },
+    {
+      "epoch": 6.4,
+      "eval_loss": 0.36003848910331726,
+      "eval_runtime": 240.8042,
+      "eval_samples_per_second": 20.764,
+      "eval_steps_per_second": 2.595,
+      "eval_wer": 0.44533608987785966,
+      "step": 4000
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 0.00019209599999999997,
+      "loss": 0.7773,
+      "step": 4500
+    },
+    {
+      "epoch": 7.2,
+      "eval_loss": 0.37241309881210327,
+      "eval_runtime": 234.1677,
+      "eval_samples_per_second": 21.352,
+      "eval_steps_per_second": 2.669,
+      "eval_wer": 0.4470226598841163,
+      "step": 4500
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 0.00018009599999999998,
+      "loss": 0.747,
+      "step": 5000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.3500821888446808,
+      "eval_runtime": 239.967,
+      "eval_samples_per_second": 20.836,
+      "eval_steps_per_second": 2.605,
+      "eval_wer": 0.41892222736051793,
+      "step": 5000
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 0.00016809599999999998,
+      "loss": 0.7279,
+      "step": 5500
+    },
+    {
+      "epoch": 8.8,
+      "eval_loss": 0.3500120937824249,
+      "eval_runtime": 236.3527,
+      "eval_samples_per_second": 21.155,
+      "eval_steps_per_second": 2.644,
+      "eval_wer": 0.42607654851609045,
+      "step": 5500
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 0.00015612,
+      "loss": 0.7153,
+      "step": 6000
+    },
+    {
+      "epoch": 9.6,
+      "eval_loss": 0.33280453085899353,
+      "eval_runtime": 237.5821,
+      "eval_samples_per_second": 21.045,
+      "eval_steps_per_second": 2.631,
+      "eval_wer": 0.3965887761486358,
+      "step": 6000
+    },
+    {
+      "epoch": 10.4,
+      "learning_rate": 0.000144144,
+      "loss": 0.7,
+      "step": 6500
+    },
+    {
+      "epoch": 10.4,
+      "eval_loss": 0.33142638206481934,
+      "eval_runtime": 238.2506,
+      "eval_samples_per_second": 20.986,
+      "eval_steps_per_second": 2.623,
+      "eval_wer": 0.3869045999836784,
+      "step": 6500
+    },
+    {
+      "epoch": 11.2,
+      "learning_rate": 0.000132168,
+      "loss": 0.6784,
+      "step": 7000
+    },
+    {
+      "epoch": 11.2,
+      "eval_loss": 0.33956480026245117,
+      "eval_runtime": 241.1503,
+      "eval_samples_per_second": 20.734,
+      "eval_steps_per_second": 2.592,
+      "eval_wer": 0.40510323440602813,
+      "step": 7000
+    },
+    {
+      "epoch": 12.0,
+      "learning_rate": 0.00012016799999999999,
+      "loss": 0.6582,
+      "step": 7500
+    },
+    {
+      "epoch": 12.0,
+      "eval_loss": 0.3236238658428192,
+      "eval_runtime": 237.147,
+      "eval_samples_per_second": 21.084,
+      "eval_steps_per_second": 2.635,
+      "eval_wer": 0.38992410434971847,
+      "step": 7500
+    },
+    {
+      "epoch": 12.8,
+      "learning_rate": 0.00010816799999999998,
+      "loss": 0.6478,
+      "step": 8000
+    },
+    {
+      "epoch": 12.8,
+      "eval_loss": 0.3263108730316162,
+      "eval_runtime": 239.702,
+      "eval_samples_per_second": 20.859,
+      "eval_steps_per_second": 2.607,
+      "eval_wer": 0.3832050270667283,
+      "step": 8000
+    },
+    {
+      "epoch": 13.6,
+      "learning_rate": 9.616799999999999e-05,
+      "loss": 0.6277,
+      "step": 8500
+    },
+    {
+      "epoch": 13.6,
+      "eval_loss": 0.3138865828514099,
+      "eval_runtime": 241.2261,
+      "eval_samples_per_second": 20.727,
+      "eval_steps_per_second": 2.591,
+      "eval_wer": 0.37692119365632054,
+      "step": 8500
+    },
+    {
+      "epoch": 14.4,
+      "learning_rate": 8.416799999999998e-05,
+      "loss": 0.6053,
+      "step": 9000
+    },
+    {
+      "epoch": 14.4,
+      "eval_loss": 0.2955300807952881,
+      "eval_runtime": 239.5196,
+      "eval_samples_per_second": 20.875,
+      "eval_steps_per_second": 2.609,
+      "eval_wer": 0.3535812409890917,
+      "step": 9000
+    },
+    {
+      "epoch": 15.2,
+      "learning_rate": 7.216799999999999e-05,
+      "loss": 0.5777,
+      "step": 9500
+    },
+    {
+      "epoch": 15.2,
+      "eval_loss": 0.27926793694496155,
+      "eval_runtime": 243.0206,
+      "eval_samples_per_second": 20.574,
+      "eval_steps_per_second": 2.572,
+      "eval_wer": 0.3413128043306765,
+      "step": 9500
+    },
+    {
+      "epoch": 16.0,
+      "learning_rate": 6.016799999999999e-05,
+      "loss": 0.5631,
+      "step": 10000
+    },
+    {
+      "epoch": 16.0,
+      "eval_loss": 0.2789314091205597,
+      "eval_runtime": 235.0589,
+      "eval_samples_per_second": 21.271,
+      "eval_steps_per_second": 2.659,
+      "eval_wer": 0.33532820108266914,
+      "step": 10000
+    },
+    {
+      "epoch": 16.8,
+      "learning_rate": 4.8168e-05,
+      "loss": 0.5446,
+      "step": 10500
+    },
+    {
+      "epoch": 16.8,
+      "eval_loss": 0.27094796299934387,
+      "eval_runtime": 244.8594,
+      "eval_samples_per_second": 20.42,
+      "eval_steps_per_second": 2.552,
+      "eval_wer": 0.32637849895269444,
+      "step": 10500
+    },
+    {
+      "epoch": 17.6,
+      "learning_rate": 3.6167999999999997e-05,
+      "loss": 0.528,
+      "step": 11000
+    },
+    {
+      "epoch": 17.6,
+      "eval_loss": 0.26928168535232544,
+      "eval_runtime": 237.8703,
+      "eval_samples_per_second": 21.02,
+      "eval_steps_per_second": 2.627,
+      "eval_wer": 0.3233861973286907,
+      "step": 11000
+    },
+    {
+      "epoch": 18.4,
+      "learning_rate": 2.4191999999999998e-05,
+      "loss": 0.5169,
+      "step": 11500
+    },
+    {
+      "epoch": 18.4,
+      "eval_loss": 0.265558123588562,
+      "eval_runtime": 240.7332,
+      "eval_samples_per_second": 20.77,
+      "eval_steps_per_second": 2.596,
+      "eval_wer": 0.31925138053915836,
+      "step": 11500
+    },
+    {
+      "epoch": 19.2,
+      "learning_rate": 1.2215999999999998e-05,
+      "loss": 0.5041,
+      "step": 12000
+    },
+    {
+      "epoch": 19.2,
+      "eval_loss": 0.2575398087501526,
+      "eval_runtime": 242.0688,
+      "eval_samples_per_second": 20.655,
+      "eval_steps_per_second": 2.582,
+      "eval_wer": 0.31024727292511084,
+      "step": 12000
+    },
+    {
+      "epoch": 20.0,
+      "learning_rate": 2.16e-07,
+      "loss": 0.4971,
+      "step": 12500
+    },
+    {
+      "epoch": 20.0,
+      "eval_loss": 0.2584246098995209,
+      "eval_runtime": 247.3331,
+      "eval_samples_per_second": 20.216,
+      "eval_steps_per_second": 2.527,
+      "eval_wer": 0.31141699083267593,
+      "step": 12500
+    },
+    {
+      "epoch": 20.0,
+      "step": 12500,
+      "total_flos": 3.751048304986411e+19,
+      "train_loss": 0.8456691796875,
+      "train_runtime": 20148.7553,
+      "train_samples_per_second": 9.926,
+      "train_steps_per_second": 0.62
+    }
+  ],
+  "max_steps": 12500,
+  "num_train_epochs": 20,
+  "total_flos": 3.751048304986411e+19,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:312fb5b03ffb7643c483c67d3f9c080ba092705560e886015616251b77d99ba2
+oid sha256:4c6b5ee3512f0a11b714e02a1d75e8b09bbdb1f199638928581165457d0f25fc
 size 3055

vocab.json

@@ -0,0 +1 @@
+{"!": 1, "\"": 2, "'": 3, "(": 4, ")": 5, ",": 6, "-": 7, ".": 8, ":": 9, ";": 10, "?": 11, "a": 12, "b": 13, "c": 14, "d": 15, "e": 16, "f": 17, "g": 18, "h": 19, "i": 20, "j": 21, "k": 22, "l": 23, "m": 24, "n": 25, "o": 26, "p": 27, "r": 28, "s": 29, "t": 30, "u": 31, "v": 32, "y": 33, "z": 34, "«": 35, "»": 36, "é": 37, "ñ": 38, "ĉ": 39, "ĝ": 40, "ĥ": 41, "ĵ": 42, "ŝ": 43, "ŭ": 44, "–": 45, "—": 46, "‘": 47, "’": 48, "“": 49, "”": 50, "„": 51, "fi": 52, "|": 0, "[UNK]": 53, "[PAD]": 54}