End of training

README.md

@@ -22,19 +22,19 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.1249
-- Precision: 0.2559
-- Recall: 0.2742
-- F1: 0.2642
-- Precision Median: 0.25
-- Recall Median: 0.2727
-- F1 Median: 0.2609
-- Precision Max: 0.4211
-- Recall Max: 0.4706
-- F1 Max: 0.4444
-- Precision Min: 0.0526
-- Recall Min: 0.05
-- F1 Min: 0.0513
+- Loss: 1.3168
+- Precision: 0.1010
+- Recall: 0.1098
+- F1: 0.1051
+- Precision Median: 0.0
+- Recall Median: 0.0
+- F1 Median: 0.0
+- Precision Max: 0.5
+- Recall Max: 0.5333
+- F1 Max: 0.5161
+- Precision Min: 0.0
+- Recall Min: 0.0
+- F1 Min: 0.0
 
 ## Model description
 
@@ -53,25 +53,24 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 8e-05
-- train_batch_size: 12
+- learning_rate: 1e-06
+- train_batch_size: 32
 - eval_batch_size: 8
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 20
-- training_steps: 25
+- lr_scheduler_warmup_steps: 200
+- training_steps: 800
 - mixed_precision_training: Native AMP
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Precision Median | Recall Median | F1 Median | Precision Max | Recall Max | F1 Max | Precision Min | Recall Min | F1 Min |
 |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:----------------:|:-------------:|:---------:|:-------------:|:----------:|:------:|:-------------:|:----------:|:------:|
-| No log        | 0.0   | 5    | 5.8029          | 0.0       | 0.0    | 0.0    | 0.0              | 0.0           | 0.0       | 0             | 0          | 0      | 0             | 0          | 0      |
-| No log        | 0.01  | 10   | 2.9216          | 0.0       | 0.0    | 0.0    | 0.0              | 0.0           | 0.0       | 0             | 0          | 0      | 0             | 0          | 0      |
-| No log        | 0.01  | 15   | 2.1863          | 0.0540    | 0.0710 | 0.0609 | 0.0476           | 0.0769        | 0.0541    | 0.2           | 0.2353     | 0.2162 | 0.0           | 0.0        | 0.0    |
-| No log        | 0.01  | 20   | 1.1833          | 0.3468    | 0.3692 | 0.3565 | 0.4118           | 0.4286        | 0.4242    | 0.8           | 0.8        | 0.8000 | 0.0           | 0.0        | 0.0    |
-| 3.5014        | 0.01  | 25   | 1.1249          | 0.2559    | 0.2742 | 0.2642 | 0.25             | 0.2727        | 0.2609    | 0.4211        | 0.4706     | 0.4444 | 0.0526        | 0.05       | 0.0513 |
+| 3.5705        | 0.32  | 200  | 2.7306          | 0.0       | 0.0    | 0.0    | 0.0              | 0.0           | 0.0       | 0             | 0          | 0      | 0             | 0          | 0      |
+| 2.0183        | 0.63  | 400  | 1.6209          | 0.0341    | 0.0373 | 0.0356 | 0.0              | 0.0           | 0.0       | 0.4375        | 0.4667     | 0.4516 | 0.0           | 0.0        | 0.0    |
+| 1.6646        | 0.95  | 600  | 1.3710          | 0.1062    | 0.1151 | 0.1103 | 0.0              | 0.0           | 0.0       | 0.5           | 0.5333     | 0.5161 | 0.0           | 0.0        | 0.0    |
+| 1.6004        | 1.27  | 800  | 1.3168          | 0.1010    | 0.1098 | 0.1051 | 0.0              | 0.0           | 0.0       | 0.5           | 0.5333     | 0.5161 | 0.0           | 0.0        | 0.0    |
 
 
 ### Framework versions

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "openai/whisper-medium",
+  "_name_or_path": "openai/whisper-tiny",
   "activation_dropout": 0.1,
   "activation_function": "gelu",
   "apply_spec_augment": false,
@@ -13,18 +13,18 @@
   ],
   "bos_token_id": 50257,
   "classifier_proj_size": 256,
-  "d_model": 1024,
-  "decoder_attention_heads": 16,
-  "decoder_ffn_dim": 4096,
+  "d_model": 384,
+  "decoder_attention_heads": 6,
+  "decoder_ffn_dim": 1536,
   "decoder_input_ids": null,
   "decoder_layerdrop": 0.0,
-  "decoder_layers": 24,
+  "decoder_layers": 4,
   "decoder_start_token_id": 50258,
   "dropout": 0.1,
-  "encoder_attention_heads": 16,
-  "encoder_ffn_dim": 4096,
+  "encoder_attention_heads": 6,
+  "encoder_ffn_dim": 1536,
   "encoder_layerdrop": 0.0,
-  "encoder_layers": 24,
+  "encoder_layers": 4,
   "eos_token_id": 50257,
   "forced_decoder_ids": null,
   "init_std": 0.02,
@@ -40,7 +40,7 @@
   "max_target_positions": 448,
   "median_filter_width": 7,
   "model_type": "whisper",
-  "num_hidden_layers": 24,
+  "num_hidden_layers": 4,
   "num_mel_bins": 80,
   "pad_token_id": 50257,
   "scale_embedding": false,

generation_config.json

@@ -1,28 +1,28 @@
 {
   "alignment_heads": [
     [
-      13,
-      15
+      2,
+      2
     ],
     [
-      15,
-      4
+      3,
+      0
     ],
     [
-      15,
-      15
+      3,
+      2
     ],
     [
-      16,
-      1
+      3,
+      3
     ],
     [
-      20,
-      0
+      3,
+      4
     ],
     [
-      23,
-      4
+      3,
+      5
     ]
   ],
   "begin_suppress_tokens": [

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f54d42f49d8b09dd8b31b9c978a06fcf8845df57976a09f0a437e88625dc530e
-size 3055671280
+oid sha256:2a58cfd5695b6da2e034f88595f4d5b0a2e8aac9f0a8d9dd68227f5d70451656
+size 151109288

runs/Feb15_06-05-10_sipl-7542-ct/events.out.tfevents.1707977115.sipl-7542-ct.1319778.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ab9bff1cf65c72f5ffe5cb8b214ed46e7a7c3834633412b5be2e9bf9cf8ac651
-size 21504
+oid sha256:8f4d87b8c190c180e9249887a9b57e46ec838dd420c3a1a830b4400edd3703fc
+size 35073

runs/Feb15_09-46-30_sipl-7542-ct/events.out.tfevents.1707990393.sipl-7542-ct.1389393.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8258ec2b78bf46bc2d7a862d79fd276c0cda3e16529c7a50066675415d80c1a7
+size 32996

runs/Feb15_09-56-22_sipl-7542-ct/events.out.tfevents.1707990986.sipl-7542-ct.1389393.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f54339a7fe03aff2d8a32382960ea502bb100b8776bb1f84d28f90f42ad51309
+size 15061

runs/Feb15_10-14-35_sipl-7542-ct/events.out.tfevents.1707992078.sipl-7542-ct.1398486.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02b06356517f3d1ca7c488f2af5ffefb093e196a1f9e632497cef97521a8b691
+size 14312

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72d81e4dee5fa4935124c935ad0735bba7504a6a2a97c04edab02cab35debb0d
+oid sha256:43481db32632992ddb738a945c08f45b8e22b8750f821072e117f72577ebc501
 size 4475