End of training

README.md

@@ -1,8 +1,6 @@
 ---
-language:
-- jpn
-license: apache-2.0
-base_model: openai/whisper-small
+license: mit
+base_model: pyannote/segmentation-3.0
 tags:
 - speaker-diarization
 - speaker-segmentation
@@ -19,17 +17,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # speaker-segmentation-fine-tuned-callhome-jpn
 
-This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the diarizers-community/callhome dataset.
+This model is a fine-tuned version of [pyannote/segmentation-3.0](https://huggingface.co/pyannote/segmentation-3.0) on the diarizers-community/callhome jpn dataset.
 It achieves the following results on the evaluation set:
-- eval_loss: 0.8263
-- eval_der: 0.2702
-- eval_false_alarm: 0.0280
-- eval_missed_detection: 0.1843
-- eval_confusion: 0.0579
-- eval_runtime: 80.4369
-- eval_samples_per_second: 8.64
-- eval_steps_per_second: 0.274
-- step: 0
+- Loss: 0.5957
+- Der: 0.1975
+- False Alarm: 0.0777
+- Missed Detection: 0.0713
+- Confusion: 0.0485
 
 ## Model description
 
@@ -54,11 +48,22 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
-- num_epochs: 5
+- num_epochs: 5.0
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss | Der    | False Alarm | Missed Detection | Confusion |
+|:-------------:|:-----:|:----:|:---------------:|:------:|:-----------:|:----------------:|:---------:|
+| 0.5998        | 1.0   | 336  | 0.6155          | 0.2067 | 0.0726      | 0.0792           | 0.0549    |
+| 0.578         | 2.0   | 672  | 0.6258          | 0.2086 | 0.0851      | 0.0691           | 0.0544    |
+| 0.5431        | 3.0   | 1008 | 0.6054          | 0.2023 | 0.0830      | 0.0689           | 0.0505    |
+| 0.5198        | 4.0   | 1344 | 0.5989          | 0.1984 | 0.0762      | 0.0729           | 0.0494    |
+| 0.5211        | 5.0   | 1680 | 0.5957          | 0.1975 | 0.0777      | 0.0713           | 0.0485    |
+
 
 ### Framework versions
 
 - Transformers 4.40.0
-- Pytorch 2.2.1+cu121
-- Datasets 2.19.0
+- Pytorch 2.2.2+cu121
+- Datasets 2.18.0
 - Tokenizers 0.19.1

emissions.csv

@@ -0,0 +1,2 @@
+timestamp,experiment_id,project_name,duration,emissions,energy_consumed,country_name,country_iso_code,region,on_cloud,cloud_provider,cloud_region
+2024-04-20T16:46:40,f247ee22-7a72-4405-80cf-efae1836a21e,codecarbon,220.32588934898376,0.0010399341439828672,0.002449301710285461,France,FRA,île-de-france,N,,