ManuD/tts_test_model

README.md

@@ -1,5 +1,6 @@
 ---
 license: mit
+base_model: microsoft/speecht5_tts
 tags:
 - generated_from_trainer
 datasets:
@@ -16,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the voxpopuli dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4636
+- Loss: 0.4596
 
 ## Model description
 
@@ -45,20 +46,21 @@ The following hyperparameters were used during training:
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 500
 - training_steps: 4000
+- mixed_precision_training: Native AMP
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.5307        | 2.26  | 1000 | 0.4842          |
-| 0.5081        | 4.52  | 2000 | 0.4712          |
-| 0.505         | 6.79  | 3000 | 0.4646          |
-| 0.4986        | 9.05  | 4000 | 0.4636          |
+| 0.5221        | 2.77  | 1000 | 0.4784          |
+| 0.506         | 5.53  | 2000 | 0.4662          |
+| 0.495         | 8.3   | 3000 | 0.4613          |
+| 0.4977        | 11.06 | 4000 | 0.4596          |
 
 
 ### Framework versions
 
-- Transformers 4.30.2
-- Pytorch 2.0.1+cu118
-- Datasets 2.13.0
-- Tokenizers 0.13.3
+- Transformers 4.38.2
+- Pytorch 2.2.1+cu121
+- Datasets 2.18.0
+- Tokenizers 0.15.2

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "decoder_start_token_id": 2,
+  "eos_token_id": 2,
+  "max_length": 1876,
+  "pad_token_id": 1,
+  "transformers_version": "4.38.2"
+}