Model save

README.md

@@ -1,66 +1,60 @@
----
-library_name: transformers
-license: apache-2.0
-base_model: google/umt5-base
-tags:
-- generated_from_trainer
-metrics:
-- wer
-model-index:
-- name: umt5-base-asr
-  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. -->
-
-# umt5-base-asr
-
-This model is a fine-tuned version of [google/umt5-base](https://huggingface.co/google/umt5-base) on an unknown dataset.
-It achieves the following results on the evaluation set:
-- Loss: 82.8125
-- Wer: 1.0003
-
-## 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: 32
-- seed: 42
-- gradient_accumulation_steps: 8
-- total_train_batch_size: 256
-- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: linear
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3
-
-### Training results
-
-| Training Loss | Epoch  | Step | Validation Loss | Wer    |
-|:-------------:|:------:|:----:|:---------------:|:------:|
-| 8.0063        | 0.9997 | 396  | 61.0312         | 1.0011 |
-| 9.9092        | 1.9972 | 792  | 61.2188         | 2.3606 |
-| 8.7422        | 2.9946 | 1188 | 82.8125         | 1.0003 |
-
-
-### Framework versions
-
-- Transformers 4.51.3
-- Pytorch 2.5.1+cu124
-- Datasets 2.17.1
-- Tokenizers 0.21.0
+---
+library_name: transformers
+license: apache-2.0
+base_model: urarik/umt5-base-asr
+tags:
+- generated_from_trainer
+model-index:
+- name: umt5-base-asr
+  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. -->
+
+# umt5-base-asr
+
+This model is a fine-tuned version of [urarik/umt5-base-asr](https://huggingface.co/urarik/umt5-base-asr) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- eval_loss: 36.1918
+- eval_wer: 1.0039
+- eval_runtime: 729.8994
+- eval_samples_per_second: 6.85
+- eval_steps_per_second: 1.713
+- epoch: 0.9997
+- step: 3168
+
+## 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: 4
+- eval_batch_size: 4
+- seed: 42
+- gradient_accumulation_steps: 8
+- total_train_batch_size: 32
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 3
+
+### Framework versions
+
+- Transformers 4.51.3
+- Pytorch 2.7.0+cu126
+- Datasets 2.17.1
+- Tokenizers 0.21.0

generation_config.json

@@ -1,6 +1,6 @@
-{
-  "decoder_start_token_id": 0,
-  "eos_token_id": 1,
-  "pad_token_id": 0,
-  "transformers_version": "4.51.3"
-}
+{
+  "decoder_start_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.51.3"
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:40c36368fe62cc98e4ea9811726d5e22293ddfeae6b46e3667532a39ac46280e
+oid sha256:c156be4b0ace3548a568605a9d5af64eb1698f40b762a453f46a30b276442115
 size 1259621296

runs/May12_19-52-33_DESKTOP-IIBMKTP/events.out.tfevents.1747050755.DESKTOP-IIBMKTP.17544.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59988a0b53b8d52d597630d1eaba5e9082a30b49f6c460b7ac2221bd34032e1f
-size 172869
+oid sha256:5e8d22549dd9eb038d578ed8cd7eed181c65f2016bbe2c375c7888fa3f18c2b7
+size 191437