End of training

README.md

@@ -1,7 +1,7 @@
 ---
 library_name: transformers
 license: mit
-base_model: voxxer/speecht5_finetuned_commonvoice_ru_translit
+base_model: microsoft/speecht5_tts
 tags:
 - generated_from_trainer
 datasets:
@@ -16,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # speecht5_finetuned_common_voice_mon
 
-This model is a fine-tuned version of [voxxer/speecht5_finetuned_commonvoice_ru_translit](https://huggingface.co/voxxer/speecht5_finetuned_commonvoice_ru_translit) on the common_voice_17_0 dataset.
+This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the common_voice_17_0 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4636
+- Loss: 0.4741
 
 ## Model description
 
@@ -53,10 +53,10 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch   | Step | Validation Loss |
 |:-------------:|:-------:|:----:|:---------------:|
-| 0.5141        | 5.9574  | 1000 | 0.4881          |
-| 0.5           | 11.9095 | 2000 | 0.4718          |
-| 0.4941        | 17.8616 | 3000 | 0.4653          |
-| 0.4873        | 23.8138 | 4000 | 0.4636          |
+| 0.5476        | 5.9574  | 1000 | 0.5062          |
+| 0.5124        | 11.9095 | 2000 | 0.4831          |
+| 0.5078        | 17.8616 | 3000 | 0.4753          |
+| 0.4994        | 23.8138 | 4000 | 0.4741          |
 
 
 ### Framework versions

generation_config.json

@@ -5,6 +5,5 @@
   "eos_token_id": 2,
   "max_length": 1876,
   "pad_token_id": 1,
-  "transformers_version": "4.48.0",
-  "use_cache": false
+  "transformers_version": "4.48.0"
 }