End of training

README.md

@@ -2,11 +2,11 @@
 language:
 - ar
 license: apache-2.0
-base_model: openai/whisper-base
+base_model: openai/whisper-small
 tags:
 - generated_from_trainer
 datasets:
-- Arbi-Houssem/Tunisian_dataset_STT-TTS
+- Arbi-Houssem/Tunisian_dataset_STT-TTS1
 metrics:
 - wer
 model-index:
@@ -16,13 +16,13 @@ model-index:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: Tunisian_dataset_STT-TTS
-      type: Arbi-Houssem/Tunisian_dataset_STT-TTS
+      name: Tunisian_dataset_STT-TTS1
+      type: Arbi-Houssem/Tunisian_dataset_STT-TTS1
       args: 'config: ar, split: test'
     metrics:
     - name: Wer
       type: wer
-      value: 123.28589909443726
+      value: 99.41634241245137
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +30,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Whisper Tunisien
 
-This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the Tunisian_dataset_STT-TTS dataset.
+This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Tunisian_dataset_STT-TTS1 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 4.4832
-- Wer: 123.2859
+- Loss: 2.9262
+- Wer: 99.4163
 
 ## Model description
 
@@ -52,8 +52,8 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 1e-05
-- train_batch_size: 16
+- learning_rate: 1e-06
+- train_batch_size: 8
 - eval_batch_size: 16
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
@@ -66,10 +66,10 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch   | Step | Validation Loss | Wer      |
 |:-------------:|:-------:|:----:|:---------------:|:--------:|
-| 0.0669        | 15.3846 | 1000 | 3.5966          | 119.1462 |
-| 0.004         | 30.7692 | 2000 | 4.2064          | 122.2510 |
-| 0.0011        | 46.1538 | 3000 | 4.4273          | 113.7128 |
-| 0.0008        | 61.5385 | 4000 | 4.4832          | 123.2859 |
+| 0.565         | 10.3093 | 1000 | 2.7181          | 99.5460  |
+| 0.2875        | 20.6186 | 2000 | 2.7486          | 106.3554 |
+| 0.1701        | 30.9278 | 3000 | 2.8744          | 103.2425 |
+| 0.1375        | 41.2371 | 4000 | 2.9262          | 99.4163  |
 
 
 ### Framework versions

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "openai/whisper-base",
+  "_name_or_path": "openai/whisper-small",
   "activation_dropout": 0.0,
   "activation_function": "gelu",
   "apply_spec_augment": false,
@@ -13,17 +13,17 @@
   ],
   "bos_token_id": 50257,
   "classifier_proj_size": 256,
-  "d_model": 512,
-  "decoder_attention_heads": 8,
-  "decoder_ffn_dim": 2048,
+  "d_model": 768,
+  "decoder_attention_heads": 12,
+  "decoder_ffn_dim": 3072,
   "decoder_layerdrop": 0.0,
-  "decoder_layers": 6,
+  "decoder_layers": 12,
   "decoder_start_token_id": 50258,
   "dropout": 0.0,
-  "encoder_attention_heads": 8,
-  "encoder_ffn_dim": 2048,
+  "encoder_attention_heads": 12,
+  "encoder_ffn_dim": 3072,
   "encoder_layerdrop": 0.0,
-  "encoder_layers": 6,
+  "encoder_layers": 12,
   "eos_token_id": 50257,
   "forced_decoder_ids": null,
   "init_std": 0.02,
@@ -39,7 +39,7 @@
   "max_target_positions": 448,
   "median_filter_width": 7,
   "model_type": "whisper",
-  "num_hidden_layers": 6,
+  "num_hidden_layers": 12,
   "num_mel_bins": 80,
   "pad_token_id": 50257,
   "scale_embedding": false,

generation_config.json

@@ -1,36 +1,44 @@
 {
   "alignment_heads": [
     [
-      3,
-      1
+      5,
+      3
     ],
     [
-      4,
-      2
+      5,
+      9
     ],
     [
-      4,
-      3
+      8,
+      0
     ],
     [
-      4,
+      8,
+      4
+    ],
+    [
+      8,
       7
     ],
     [
-      5,
-      1
+      8,
+      8
     ],
     [
-      5,
-      2
+      9,
+      0
     ],
     [
-      5,
-      4
+      9,
+      7
     ],
     [
-      5,
-      6
+      9,
+      9
+    ],
+    [
+      10,
+      5
     ]
   ],
   "begin_suppress_tokens": [

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a8d3836f14f21543c48627342fda394df5c8132c69d7051e764106af181dff0a
-size 290403936
+oid sha256:ece54d87ddf30cc6133edc4510dd41bb73d787158430c61cab6e9427e6521141
+size 966995080

runs/Jun09_02-20-59_cxrocket-SERVER/events.out.tfevents.1717896060.cxrocket-SERVER.218772.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59fcb9ee8acb8ff198fecca073cd3922a17cfe8aa57fd8f9143d2bbb365357e7
+size 41074

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25c192f2b39c838ca064cc78fdd0cdcd3008fde188dc1b0f1a5a8b66076d4ee0
+oid sha256:7c8966eba90416ac68c6f5f018f21d8f0770433b71c36d0201111b1c487bbca7
 size 5304