commit da9e809435ea7efb66d5afbd2ad590104388df5f
Author: ModelHub XC <noreply@modelhub.org.cn>
Date:   Wed May 27 04:48:16 2026 +0800

    初始化项目，由ModelHub XC社区提供模型
    
    Model: Kittipong/wav2vec2-th-vocal-domain
    Source: Original Platform

diff --git a/.gitattributes b/.gitattributes
new file mode 100644
index 0000000..ac481c8
--- /dev/null
+++ b/.gitattributes
@@ -0,0 +1,27 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zstandard filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..685f5c8
--- /dev/null
+++ b/README.md
@@ -0,0 +1,3 @@
+---
+license: cc-by-sa-4.0
+---
diff --git a/config.json b/config.json
new file mode 100644
index 0000000..e49f699
--- /dev/null
+++ b/config.json
@@ -0,0 +1,85 @@
+{
+  "_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,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "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,
+  "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": "layer",
+  "feat_proj_dropout": 0.0,
+  "feat_quantizer_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_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": 24,
+  "num_negatives": 100,
+  "pad_token_id": 69,
+  "proj_codevector_dim": 768,
+  "torch_dtype": "float32",
+  "transformers_version": "4.9.1",
+  "vocab_size": 70
+}
diff --git a/eval.py b/eval.py
new file mode 100644
index 0000000..f0af998
--- /dev/null
+++ b/eval.py
@@ -0,0 +1,154 @@
+#!/usr/bin/env python3
+import argparse
+import re
+from typing import Dict
+
+from datasets import Audio, Dataset, load_dataset, load_metric
+
+from transformers import AutoFeatureExtractor, pipeline
+
+from pythainlp.tokenize import word_tokenize, syllable_tokenize
+from deepcut import tokenize as deepcut_word_tokenize
+from functools import partial
+
+
+def log_results(result: Dataset, args: Dict[str, str]):
+    """DO NOT CHANGE. This function computes and logs the result metrics."""
+
+    log_outputs = args.log_outputs
+    dataset_id = "_".join(args.dataset.split("/") + [args.config, args.split])
+
+    # load metric
+    wer = load_metric("wer")
+    cer = load_metric("cer")
+
+    # compute metrics
+    wer_result = wer.compute(references=result["target"], predictions=result["prediction"])
+    cer_result = cer.compute(references=result["target"], predictions=result["prediction"])
+
+    # print & log results
+    result_str = f"WER: {wer_result}\n" f"CER: {cer_result}"
+    print(result_str)
+
+    with open(f"robust-speech-event/{dataset_id}_eval_results_{args.thai_tokenizer}.txt", "w") as f:
+        f.write(result_str)
+
+    # log all results in text file. Possibly interesting for analysis
+    if log_outputs is not None:
+        pred_file = f"robust-speech-event/log_{dataset_id}_predictions_{args.thai_tokenizer}.txt"
+        target_file = f"robust-speech-event/log_{dataset_id}_targets_{args.thai_tokenizer}.txt"
+
+        with open(pred_file, "w") as p, open(target_file, "w") as t:
+
+            # mapping function to write output
+            def write_to_file(batch, i):
+                p.write(f"{i}" + "\n")
+                p.write(batch["prediction"] + "\n")
+                t.write(f"{i}" + "\n")
+                t.write(batch["target"] + "\n")
+
+            result.map(write_to_file, with_indices=True)
+
+
+def normalize_text(text: str, tok_func) -> str:
+    """DO ADAPT FOR YOUR USE CASE. this function normalizes the target text."""
+
+    chars_to_ignore_regex = '[,?.!\-\;\:"“%‘”�—’…–]'  # noqa: W605 IMPORTANT: this should correspond to the chars that were ignored during training
+
+    text = re.sub(chars_to_ignore_regex, "", text.lower())
+
+    # In addition, we can normalize the target text, e.g. removing new lines characters etc...
+    # note that order is important here!
+    token_sequences_to_ignore = ["\n\n", "\n", "   ", "  "]
+
+    for t in token_sequences_to_ignore:
+        text = " ".join(text.split(t))
+
+    #thai tokenize
+    text = " ".join(tok_func(text))
+    
+    return text
+
+def retokenize(text:str, tok_func) -> str:
+    """tokenize and rejoin prediction outputs without cleaning"""
+    return " ".join(tok_func("".join(text.split())))
+
+
+def main(args):
+    # load dataset
+    dataset = load_dataset(args.dataset, args.config, split=args.split, use_auth_token=True)
+
+    # for testing: only process the first two examples as a test
+    dataset = dataset.select(range(10))
+
+    # load processor
+    feature_extractor = AutoFeatureExtractor.from_pretrained(args.model_id)
+    sampling_rate = feature_extractor.sampling_rate
+
+    # resample audio
+    dataset = dataset.cast_column("audio", Audio(sampling_rate=sampling_rate))
+
+    # load eval pipeline
+    asr = pipeline("automatic-speech-recognition", model=args.model_id)
+
+    #select tokenizer
+    if args.thai_tokenizer=='deepcut':
+        tok_func = deepcut_word_tokenize
+    elif args.thai_tokenizer=='newmm':
+        tok_func = word_tokenize
+    elif args.thai_tokenizer=='syllable':
+        tok_func = syllable_tokenize
+    else:
+        tok_func = lambda x: x.replace(' ','')
+    
+    # map function to decode audio
+    def map_to_pred(batch, tok_func):
+        prediction = asr(
+            batch["audio"]["array"], chunk_length_s=args.chunk_length_s, stride_length_s=args.stride_length_s
+        )
+
+        batch["prediction"] = retokenize(prediction["text"], tok_func)
+        batch["target"] = normalize_text(batch["sentence"], tok_func)
+        return batch
+
+    # run inference on all examples
+    result = dataset.map(partial(map_to_pred, tok_func=tok_func), 
+                         remove_columns=dataset.column_names)
+
+    # compute and log_results
+    # do not change function below
+    log_results(result, args)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--model_id", type=str, required=True, help="Model identifier. Should be loadable with 🤗 Transformers"
+    )
+    parser.add_argument(
+        "--thai_tokenizer", type=str, default="newmm",
+        required=True, help="newmm, syllable, or deepcut; if not specified, remove all spaces (used for CER calculation)"
+    )    
+    parser.add_argument(
+        "--dataset",
+        type=str,
+        required=True,
+        help="Dataset name to evaluate the `model_id`. Should be loadable with 🤗 Datasets",
+    )
+    parser.add_argument(
+        "--config", type=str, required=True, help="Config of the dataset. *E.g.* `'en'`  for Common Voice"
+    )
+    parser.add_argument("--split", type=str, required=True, help="Split of the dataset. *E.g.* `'test'`")
+    parser.add_argument(
+        "--chunk_length_s", type=float, default=None, help="Chunk length in seconds. Defaults to 5 seconds."
+    )
+    parser.add_argument(
+        "--stride_length_s", type=float, default=None, help="Stride of the audio chunks. Defaults to 1 second."
+    )
+    parser.add_argument(
+        "--log_outputs", action="store_true", help="If defined, write outputs to log file for analysis."
+    )
+    args = parser.parse_args()
+
+    main(args)
\ No newline at end of file
diff --git a/optimizer-002.pt b/optimizer-002.pt
new file mode 100644
index 0000000..ed65d80
--- /dev/null
+++ b/optimizer-002.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63480cd0bddf763bbfa016f364b5af6262bbd4584e1e3f1223bdf042feaf7080
+size 2490632977
diff --git a/preprocessor_config.json b/preprocessor_config.json
new file mode 100644
index 0000000..a0b7227
--- /dev/null
+++ b/preprocessor_config.json
@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}
diff --git a/pytorch_model.bin b/pytorch_model.bin
new file mode 100644
index 0000000..2a368c6
--- /dev/null
+++ b/pytorch_model.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dbb3a749c3f6e08fe2af7227d4857723fd1e525157b79ccb0dba255a69fdce8
+size 1262210673
diff --git a/rng_state.pth b/rng_state.pth
new file mode 100644
index 0000000..0742e20
--- /dev/null
+++ b/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e9b980149921cf7dc2fe91f225c1c3f931cc22ff40d7e30639d7236b390d622
+size 14567
diff --git a/scheduler.pt b/scheduler.pt
new file mode 100644
index 0000000..e440370
--- /dev/null
+++ b/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b9ed0945bc75b9b51a6a34f17bea3f2eca616815e76668e99905ed244654b22
+size 623
diff --git a/special_tokens_map.json b/special_tokens_map.json
new file mode 100644
index 0000000..9abf719
--- /dev/null
+++ b/special_tokens_map.json
@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}
\ No newline at end of file
diff --git a/tokenizer_config.json b/tokenizer_config.json
new file mode 100644
index 0000000..f5118ab
--- /dev/null
+++ b/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": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
\ No newline at end of file
diff --git a/trainer_state.json b/trainer_state.json
new file mode 100644
index 0000000..d776f24
--- /dev/null
+++ b/trainer_state.json
@@ -0,0 +1,616 @@
+{
+  "best_metric": 0.472552783109405,
+  "best_model_checkpoint": "/content/drive/MyDrive/new_dataset/wav2vec2-large-xlsr-53-thai-finetune/checkpoint-500",
+  "epoch": 99.98765432098766,
+  "global_step": 2000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 2.49,
+      "learning_rate": 0.0001,
+      "loss": 1.9911,
+      "step": 50
+    },
+    {
+      "epoch": 2.49,
+      "eval_loss": 1.2750070095062256,
+      "eval_runtime": 34.892,
+      "eval_samples_per_second": 12.238,
+      "eval_steps_per_second": 0.774,
+      "eval_wer": 0.6767754318618042,
+      "step": 50
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 9.743589743589744e-05,
+      "loss": 1.2107,
+      "step": 100
+    },
+    {
+      "epoch": 4.99,
+      "eval_loss": 1.1088056564331055,
+      "eval_runtime": 34.7796,
+      "eval_samples_per_second": 12.277,
+      "eval_steps_per_second": 0.776,
+      "eval_wer": 0.5927063339731286,
+      "step": 100
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 9.487179487179487e-05,
+      "loss": 0.9686,
+      "step": 150
+    },
+    {
+      "epoch": 7.49,
+      "eval_loss": 1.0476980209350586,
+      "eval_runtime": 34.6203,
+      "eval_samples_per_second": 12.334,
+      "eval_steps_per_second": 0.78,
+      "eval_wer": 0.5712092130518234,
+      "step": 150
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 9.230769230769232e-05,
+      "loss": 0.7504,
+      "step": 200
+    },
+    {
+      "epoch": 9.99,
+      "eval_loss": 1.0640665292739868,
+      "eval_runtime": 34.843,
+      "eval_samples_per_second": 12.255,
+      "eval_steps_per_second": 0.775,
+      "eval_wer": 0.5504798464491363,
+      "step": 200
+    },
+    {
+      "epoch": 12.49,
+      "learning_rate": 8.974358974358975e-05,
+      "loss": 0.6352,
+      "step": 250
+    },
+    {
+      "epoch": 12.49,
+      "eval_loss": 1.095747709274292,
+      "eval_runtime": 33.9281,
+      "eval_samples_per_second": 12.585,
+      "eval_steps_per_second": 0.796,
+      "eval_wer": 0.5324376199616123,
+      "step": 250
+    },
+    {
+      "epoch": 14.99,
+      "learning_rate": 8.717948717948718e-05,
+      "loss": 0.5313,
+      "step": 300
+    },
+    {
+      "epoch": 14.99,
+      "eval_loss": 1.0491594076156616,
+      "eval_runtime": 33.7693,
+      "eval_samples_per_second": 12.645,
+      "eval_steps_per_second": 0.8,
+      "eval_wer": 0.5163147792706334,
+      "step": 300
+    },
+    {
+      "epoch": 17.49,
+      "learning_rate": 8.461538461538461e-05,
+      "loss": 0.4461,
+      "step": 350
+    },
+    {
+      "epoch": 17.49,
+      "eval_loss": 1.0721960067749023,
+      "eval_runtime": 33.9128,
+      "eval_samples_per_second": 12.591,
+      "eval_steps_per_second": 0.796,
+      "eval_wer": 0.5124760076775432,
+      "step": 350
+    },
+    {
+      "epoch": 19.99,
+      "learning_rate": 8.205128205128205e-05,
+      "loss": 0.4094,
+      "step": 400
+    },
+    {
+      "epoch": 19.99,
+      "eval_loss": 1.057926058769226,
+      "eval_runtime": 34.0347,
+      "eval_samples_per_second": 12.546,
+      "eval_steps_per_second": 0.793,
+      "eval_wer": 0.49750479846449136,
+      "step": 400
+    },
+    {
+      "epoch": 22.49,
+      "learning_rate": 7.948717948717948e-05,
+      "loss": 0.3467,
+      "step": 450
+    },
+    {
+      "epoch": 22.49,
+      "eval_loss": 1.0208101272583008,
+      "eval_runtime": 34.3752,
+      "eval_samples_per_second": 12.422,
+      "eval_steps_per_second": 0.785,
+      "eval_wer": 0.491362763915547,
+      "step": 450
+    },
+    {
+      "epoch": 24.99,
+      "learning_rate": 7.692307692307693e-05,
+      "loss": 0.3195,
+      "step": 500
+    },
+    {
+      "epoch": 24.99,
+      "eval_loss": 1.0337833166122437,
+      "eval_runtime": 33.8794,
+      "eval_samples_per_second": 12.604,
+      "eval_steps_per_second": 0.797,
+      "eval_wer": 0.472552783109405,
+      "step": 500
+    },
+    {
+      "epoch": 27.49,
+      "learning_rate": 7.435897435897436e-05,
+      "loss": 0.3005,
+      "step": 550
+    },
+    {
+      "epoch": 27.49,
+      "eval_loss": 1.0594605207443237,
+      "eval_runtime": 33.9955,
+      "eval_samples_per_second": 12.56,
+      "eval_steps_per_second": 0.794,
+      "eval_wer": 0.47946257197696734,
+      "step": 550
+    },
+    {
+      "epoch": 29.99,
+      "learning_rate": 7.17948717948718e-05,
+      "loss": 0.2933,
+      "step": 600
+    },
+    {
+      "epoch": 29.99,
+      "eval_loss": 1.017477035522461,
+      "eval_runtime": 33.8622,
+      "eval_samples_per_second": 12.61,
+      "eval_steps_per_second": 0.797,
+      "eval_wer": 0.472936660268714,
+      "step": 600
+    },
+    {
+      "epoch": 32.49,
+      "learning_rate": 6.923076923076924e-05,
+      "loss": 0.2601,
+      "step": 650
+    },
+    {
+      "epoch": 32.49,
+      "eval_loss": 1.099133014678955,
+      "eval_runtime": 34.4437,
+      "eval_samples_per_second": 12.397,
+      "eval_steps_per_second": 0.784,
+      "eval_wer": 0.4652591170825336,
+      "step": 650
+    },
+    {
+      "epoch": 34.99,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 0.2226,
+      "step": 700
+    },
+    {
+      "epoch": 34.99,
+      "eval_loss": 1.1290050745010376,
+      "eval_runtime": 34.6356,
+      "eval_samples_per_second": 12.328,
+      "eval_steps_per_second": 0.78,
+      "eval_wer": 0.47063339731285986,
+      "step": 700
+    },
+    {
+      "epoch": 37.49,
+      "learning_rate": 6.410256410256412e-05,
+      "loss": 0.2262,
+      "step": 750
+    },
+    {
+      "epoch": 37.49,
+      "eval_loss": 1.0954631567001343,
+      "eval_runtime": 35.3666,
+      "eval_samples_per_second": 12.074,
+      "eval_steps_per_second": 0.763,
+      "eval_wer": 0.45681381957773515,
+      "step": 750
+    },
+    {
+      "epoch": 39.99,
+      "learning_rate": 6.153846153846155e-05,
+      "loss": 0.2236,
+      "step": 800
+    },
+    {
+      "epoch": 39.99,
+      "eval_loss": 1.1119202375411987,
+      "eval_runtime": 35.0365,
+      "eval_samples_per_second": 12.187,
+      "eval_steps_per_second": 0.771,
+      "eval_wer": 0.4491362763915547,
+      "step": 800
+    },
+    {
+      "epoch": 42.49,
+      "learning_rate": 5.897435897435898e-05,
+      "loss": 0.2029,
+      "step": 850
+    },
+    {
+      "epoch": 42.49,
+      "eval_loss": 1.1375640630722046,
+      "eval_runtime": 35.2676,
+      "eval_samples_per_second": 12.107,
+      "eval_steps_per_second": 0.766,
+      "eval_wer": 0.45220729366602685,
+      "step": 850
+    },
+    {
+      "epoch": 44.99,
+      "learning_rate": 5.6410256410256414e-05,
+      "loss": 0.1876,
+      "step": 900
+    },
+    {
+      "epoch": 44.99,
+      "eval_loss": 1.142003059387207,
+      "eval_runtime": 35.6836,
+      "eval_samples_per_second": 11.966,
+      "eval_steps_per_second": 0.757,
+      "eval_wer": 0.4476007677543186,
+      "step": 900
+    },
+    {
+      "epoch": 47.49,
+      "learning_rate": 5.384615384615385e-05,
+      "loss": 0.1733,
+      "step": 950
+    },
+    {
+      "epoch": 47.49,
+      "eval_loss": 1.1137712001800537,
+      "eval_runtime": 34.9014,
+      "eval_samples_per_second": 12.234,
+      "eval_steps_per_second": 0.774,
+      "eval_wer": 0.44337811900191937,
+      "step": 950
+    },
+    {
+      "epoch": 49.99,
+      "learning_rate": 5.128205128205128e-05,
+      "loss": 0.161,
+      "step": 1000
+    },
+    {
+      "epoch": 49.99,
+      "eval_loss": 1.195468783378601,
+      "eval_runtime": 35.1209,
+      "eval_samples_per_second": 12.158,
+      "eval_steps_per_second": 0.769,
+      "eval_wer": 0.44798464491362766,
+      "step": 1000
+    },
+    {
+      "epoch": 52.49,
+      "learning_rate": 4.871794871794872e-05,
+      "loss": 0.1661,
+      "step": 1050
+    },
+    {
+      "epoch": 52.49,
+      "eval_loss": 1.1598896980285645,
+      "eval_runtime": 34.8413,
+      "eval_samples_per_second": 12.256,
+      "eval_steps_per_second": 0.775,
+      "eval_wer": 0.44606525911708256,
+      "step": 1050
+    },
+    {
+      "epoch": 54.99,
+      "learning_rate": 4.615384615384616e-05,
+      "loss": 0.1533,
+      "step": 1100
+    },
+    {
+      "epoch": 54.99,
+      "eval_loss": 1.143904447555542,
+      "eval_runtime": 34.3532,
+      "eval_samples_per_second": 12.43,
+      "eval_steps_per_second": 0.786,
+      "eval_wer": 0.4383877159309021,
+      "step": 1100
+    },
+    {
+      "epoch": 57.49,
+      "learning_rate": 4.358974358974359e-05,
+      "loss": 0.1565,
+      "step": 1150
+    },
+    {
+      "epoch": 57.49,
+      "eval_loss": 1.1354175806045532,
+      "eval_runtime": 34.4974,
+      "eval_samples_per_second": 12.378,
+      "eval_steps_per_second": 0.783,
+      "eval_wer": 0.4418426103646833,
+      "step": 1150
+    },
+    {
+      "epoch": 59.99,
+      "learning_rate": 4.1025641025641023e-05,
+      "loss": 0.1476,
+      "step": 1200
+    },
+    {
+      "epoch": 59.99,
+      "eval_loss": 1.1580840349197388,
+      "eval_runtime": 34.4611,
+      "eval_samples_per_second": 12.391,
+      "eval_steps_per_second": 0.783,
+      "eval_wer": 0.4410748560460653,
+      "step": 1200
+    },
+    {
+      "epoch": 62.49,
+      "learning_rate": 3.846153846153846e-05,
+      "loss": 0.1378,
+      "step": 1250
+    },
+    {
+      "epoch": 62.49,
+      "eval_loss": 1.1266977787017822,
+      "eval_runtime": 34.2613,
+      "eval_samples_per_second": 12.463,
+      "eval_steps_per_second": 0.788,
+      "eval_wer": 0.4303262955854127,
+      "step": 1250
+    },
+    {
+      "epoch": 64.99,
+      "learning_rate": 3.58974358974359e-05,
+      "loss": 0.1214,
+      "step": 1300
+    },
+    {
+      "epoch": 64.99,
+      "eval_loss": 1.165766716003418,
+      "eval_runtime": 33.9904,
+      "eval_samples_per_second": 12.562,
+      "eval_steps_per_second": 0.794,
+      "eval_wer": 0.42879078694817657,
+      "step": 1300
+    },
+    {
+      "epoch": 67.49,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.133,
+      "step": 1350
+    },
+    {
+      "epoch": 67.49,
+      "eval_loss": 1.137351632118225,
+      "eval_runtime": 34.008,
+      "eval_samples_per_second": 12.556,
+      "eval_steps_per_second": 0.794,
+      "eval_wer": 0.43339731285988486,
+      "step": 1350
+    },
+    {
+      "epoch": 69.99,
+      "learning_rate": 3.0769230769230774e-05,
+      "loss": 0.1147,
+      "step": 1400
+    },
+    {
+      "epoch": 69.99,
+      "eval_loss": 1.1521999835968018,
+      "eval_runtime": 33.8683,
+      "eval_samples_per_second": 12.608,
+      "eval_steps_per_second": 0.797,
+      "eval_wer": 0.4245681381957774,
+      "step": 1400
+    },
+    {
+      "epoch": 72.49,
+      "learning_rate": 2.8205128205128207e-05,
+      "loss": 0.125,
+      "step": 1450
+    },
+    {
+      "epoch": 72.49,
+      "eval_loss": 1.1379369497299194,
+      "eval_runtime": 34.1582,
+      "eval_samples_per_second": 12.501,
+      "eval_steps_per_second": 0.79,
+      "eval_wer": 0.43570057581573896,
+      "step": 1450
+    },
+    {
+      "epoch": 74.99,
+      "learning_rate": 2.564102564102564e-05,
+      "loss": 0.1189,
+      "step": 1500
+    },
+    {
+      "epoch": 74.99,
+      "eval_loss": 1.1502233743667603,
+      "eval_runtime": 33.9539,
+      "eval_samples_per_second": 12.576,
+      "eval_steps_per_second": 0.795,
+      "eval_wer": 0.4284069097888676,
+      "step": 1500
+    },
+    {
+      "epoch": 77.49,
+      "learning_rate": 2.307692307692308e-05,
+      "loss": 0.1122,
+      "step": 1550
+    },
+    {
+      "epoch": 77.49,
+      "eval_loss": 1.1575168371200562,
+      "eval_runtime": 34.1792,
+      "eval_samples_per_second": 12.493,
+      "eval_steps_per_second": 0.79,
+      "eval_wer": 0.4314779270633397,
+      "step": 1550
+    },
+    {
+      "epoch": 79.99,
+      "learning_rate": 2.0512820512820512e-05,
+      "loss": 0.1235,
+      "step": 1600
+    },
+    {
+      "epoch": 79.99,
+      "eval_loss": 1.142774224281311,
+      "eval_runtime": 33.8121,
+      "eval_samples_per_second": 12.629,
+      "eval_steps_per_second": 0.799,
+      "eval_wer": 0.42610364683301344,
+      "step": 1600
+    },
+    {
+      "epoch": 82.49,
+      "learning_rate": 1.794871794871795e-05,
+      "loss": 0.1092,
+      "step": 1650
+    },
+    {
+      "epoch": 82.49,
+      "eval_loss": 1.1530485153198242,
+      "eval_runtime": 33.6852,
+      "eval_samples_per_second": 12.676,
+      "eval_steps_per_second": 0.802,
+      "eval_wer": 0.4341650671785029,
+      "step": 1650
+    },
+    {
+      "epoch": 84.99,
+      "learning_rate": 1.5384615384615387e-05,
+      "loss": 0.1185,
+      "step": 1700
+    },
+    {
+      "epoch": 84.99,
+      "eval_loss": 1.1284310817718506,
+      "eval_runtime": 33.8486,
+      "eval_samples_per_second": 12.615,
+      "eval_steps_per_second": 0.798,
+      "eval_wer": 0.42994241842610365,
+      "step": 1700
+    },
+    {
+      "epoch": 87.49,
+      "learning_rate": 1.282051282051282e-05,
+      "loss": 0.1048,
+      "step": 1750
+    },
+    {
+      "epoch": 87.49,
+      "eval_loss": 1.1434178352355957,
+      "eval_runtime": 33.9335,
+      "eval_samples_per_second": 12.583,
+      "eval_steps_per_second": 0.796,
+      "eval_wer": 0.4276391554702495,
+      "step": 1750
+    },
+    {
+      "epoch": 89.99,
+      "learning_rate": 1.0256410256410256e-05,
+      "loss": 0.1038,
+      "step": 1800
+    },
+    {
+      "epoch": 89.99,
+      "eval_loss": 1.1609505414962769,
+      "eval_runtime": 34.1384,
+      "eval_samples_per_second": 12.508,
+      "eval_steps_per_second": 0.791,
+      "eval_wer": 0.42418426103646834,
+      "step": 1800
+    },
+    {
+      "epoch": 92.49,
+      "learning_rate": 7.692307692307694e-06,
+      "loss": 0.1073,
+      "step": 1850
+    },
+    {
+      "epoch": 92.49,
+      "eval_loss": 1.1562278270721436,
+      "eval_runtime": 33.8455,
+      "eval_samples_per_second": 12.616,
+      "eval_steps_per_second": 0.798,
+      "eval_wer": 0.4238003838771593,
+      "step": 1850
+    },
+    {
+      "epoch": 94.99,
+      "learning_rate": 5.128205128205128e-06,
+      "loss": 0.1055,
+      "step": 1900
+    },
+    {
+      "epoch": 94.99,
+      "eval_loss": 1.1589001417160034,
+      "eval_runtime": 34.1041,
+      "eval_samples_per_second": 12.52,
+      "eval_steps_per_second": 0.792,
+      "eval_wer": 0.4214971209213052,
+      "step": 1900
+    },
+    {
+      "epoch": 97.49,
+      "learning_rate": 2.564102564102564e-06,
+      "loss": 0.1133,
+      "step": 1950
+    },
+    {
+      "epoch": 97.49,
+      "eval_loss": 1.1536645889282227,
+      "eval_runtime": 33.9701,
+      "eval_samples_per_second": 12.57,
+      "eval_steps_per_second": 0.795,
+      "eval_wer": 0.4238003838771593,
+      "step": 1950
+    },
+    {
+      "epoch": 99.99,
+      "learning_rate": 0.0,
+      "loss": 0.0992,
+      "step": 2000
+    },
+    {
+      "epoch": 99.99,
+      "eval_loss": 1.1553977727890015,
+      "eval_runtime": 33.8655,
+      "eval_samples_per_second": 12.609,
+      "eval_steps_per_second": 0.797,
+      "eval_wer": 0.42418426103646834,
+      "step": 2000
+    }
+  ],
+  "max_steps": 2000,
+  "num_train_epochs": 100,
+  "total_flos": 1.6479245346890066e+19,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/training_args.bin b/training_args.bin
new file mode 100644
index 0000000..8934fac
--- /dev/null
+++ b/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a007a4ec524a3af0145655a9608fca61329594a670f7bbbd2bb694f2d8648c8
+size 2799
diff --git a/vocab.json b/vocab.json
new file mode 100644
index 0000000..0046279
--- /dev/null
+++ b/vocab.json
@@ -0,0 +1 @@
+{"ฑ": 0, "ๅ": 1, "ก": 2, "ง": 3, "ฒ": 4, "ะ": 5, "๊": 6, "้": 7, "ฌ": 8, "ซ": 9, "ด": 10, "ฯ": 11, "ใ": 12, "ึ": 13, "ญ": 14, "่": 15, "า": 16, "ฤ": 17, "๋": 18, "อ": 19, "ฬ": 20, "ท": 21, "โ": 22, "ภ": 23, "ย": 24, "็": 25, "ล": 26, "ุ": 27, "เ": 28, "ฮ": 29, "ฝ": 30, "ป": 31, "ี": 32, "บ": 33, "ฐ": 34, "ต": 35, "ถ": 36, "ศ": 37, "ฟ": 38, "ณ": 39, "ห": 40, "ร": 41, "พ": 43, "ฆ": 44, "ั": 45, "ค": 46, "ว": 47, "ฏ": 48, "จ": 49, "แ": 50, "ม": 51, "ฎ": 52, "ฉ": 53, "์": 54, "ษ": 55, "ำ": 56, "ผ": 57, "ข": 58, "ไ": 59, "ู": 60, "ื": 61, "น": 62, "ช": 63, "ิ": 64, "ธ": 65, "ฃ": 66, "ส": 67, "|": 42, "[UNK]": 68, "[PAD]": 69}
\ No newline at end of file