Initial commit

README.md

@@ -0,0 +1,52 @@
+---
+language: tr
+datasets:
+- common_voice
+metrics:
+- wer
+tags:
+- audio
+- automatic-speech-recognition
+- speech
+
+license: apache-2.0
+model-index:
+- name: XLSR Wav2Vec2 Turkish by Davut Emre TASAR
+  results:
+  - task: 
+      name: Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: Common Voice tr
+      type: common_voice
+      args: tr 
+    metrics:
+       - name: Test WER
+         type: wer
+         
+---
+
+# wav2vec-tr-lite-AG
+
+## Usage
+
+The model can be used directly (without a language model) as follows:
+
+```python
+import torch
+import torchaudio
+from datasets import load_dataset
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+
+test_dataset = load_dataset("common_voice", "tr", split="test[:2%]") 
+
+processor = Wav2Vec2Processor.from_pretrained("emre/wav2vec-tr-lite-AG")
+model = Wav2Vec2ForCTC.from_pretrained("emre/wav2vec-tr-lite-AG")
+
+resampler = torchaudio.transforms.Resample(48_000, 16_000)
+
+**Test Result**: 27.30 %
+
+
+ [here](https://adresgezgini.com)
+

config.json

@@ -0,0 +1,76 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.0,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "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": "mean",
+  "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": 0.0,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "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_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "pad_token_id": 39,
+  "transformers_version": "4.4.0",
+  "vocab_size": 40
+}

preprocessor_config.json

@@ -0,0 +1,8 @@
+{
+  "do_normalize": true,
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:721f3731148ec50eb9a3884b730f28079078ce2ed4fa90dbcb1d8a8fd5ef4633
+size 1262097815

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fbdd95cbb152ba0c0f75f3312676c3c1de2346ccbe8f408f9381dbfd9f76c8a
+size 623

trainer_state.json

@@ -0,0 +1,128 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 29.357798165137616,
+  "global_step": 3200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 3.67,
+      "learning_rate": 0.00023999999999999998,
+      "loss": 5.0456,
+      "step": 400
+    },
+    {
+      "epoch": 3.67,
+      "eval_loss": 2.805786609649658,
+      "eval_runtime": 214.876,
+      "eval_samples_per_second": 7.665,
+      "eval_wer": 1.0,
+      "step": 400
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 0.0002675090252707581,
+      "loss": 0.8437,
+      "step": 800
+    },
+    {
+      "epoch": 7.34,
+      "eval_loss": 0.437711238861084,
+      "eval_runtime": 217.8287,
+      "eval_samples_per_second": 7.561,
+      "eval_wer": 0.5915636809314677,
+      "step": 800
+    },
+    {
+      "epoch": 11.01,
+      "learning_rate": 0.00022418772563176893,
+      "loss": 0.2449,
+      "step": 1200
+    },
+    {
+      "epoch": 11.01,
+      "eval_loss": 0.3937138020992279,
+      "eval_runtime": 222.9363,
+      "eval_samples_per_second": 7.388,
+      "eval_wer": 0.5292615667449698,
+      "step": 1200
+    },
+    {
+      "epoch": 14.68,
+      "learning_rate": 0.00018086642599277977,
+      "loss": 0.1486,
+      "step": 1600
+    },
+    {
+      "epoch": 14.68,
+      "eval_loss": 0.411156564950943,
+      "eval_runtime": 224.0258,
+      "eval_samples_per_second": 7.352,
+      "eval_wer": 0.5094474517413952,
+      "step": 1600
+    },
+    {
+      "epoch": 18.35,
+      "learning_rate": 0.00013754512635379062,
+      "loss": 0.1077,
+      "step": 2000
+    },
+    {
+      "epoch": 18.35,
+      "eval_loss": 0.41311565041542053,
+      "eval_runtime": 224.2133,
+      "eval_samples_per_second": 7.346,
+      "eval_wer": 0.49474006740884485,
+      "step": 2000
+    },
+    {
+      "epoch": 22.02,
+      "learning_rate": 9.422382671480144e-05,
+      "loss": 0.0791,
+      "step": 2400
+    },
+    {
+      "epoch": 22.02,
+      "eval_loss": 0.43790802359580994,
+      "eval_runtime": 224.8995,
+      "eval_samples_per_second": 7.323,
+      "eval_wer": 0.48707997140230824,
+      "step": 2400
+    },
+    {
+      "epoch": 25.69,
+      "learning_rate": 5.090252707581227e-05,
+      "loss": 0.0722,
+      "step": 2800
+    },
+    {
+      "epoch": 25.69,
+      "eval_loss": 0.4185827076435089,
+      "eval_runtime": 222.949,
+      "eval_samples_per_second": 7.387,
+      "eval_wer": 0.46798079869267695,
+      "step": 2800
+    },
+    {
+      "epoch": 29.36,
+      "learning_rate": 7.581227436823104e-06,
+      "loss": 0.0601,
+      "step": 3200
+    },
+    {
+      "epoch": 29.36,
+      "eval_loss": 0.41515499353408813,
+      "eval_runtime": 225.7543,
+      "eval_samples_per_second": 7.296,
+      "eval_wer": 0.4620569911142886,
+      "step": 3200
+    }
+  ],
+  "max_steps": 3270,
+  "num_train_epochs": 30,
+  "total_flos": 1.260196128535726e+19,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80a19ab79bf56c437d7195ed88bedebbf61fec14af73bb3010f0b909dc031e04
+size 2287

vocab.json

@@ -0,0 +1 @@
+{"x": 0, "d": 1, "\u015f": 2, "\u0131": 3, "c": 4, "g": 5, "i": 6, "b": 8, "f": 9, "k": 10, "r": 11, "\u00eb": 12, "j": 13, "u": 14, "s": 15, "y": 16, "\u00f6": 17, "\u00e2": 18, "'": 19, "\u011f": 20, "n": 21, "l": 22, "e": 23, "\u00fc": 24, "p": 25, "a": 26, "\u00ee": 27, "o": 28, "w": 29, "m": 30, "v": 31, "z": 32, "q": 33, "\u00e7": 34, "t": 35, "\u0307": 36, "h": 37, "|": 7, "[UNK]": 38, "[PAD]": 39}