Upload 5 files

README.md

@@ -0,0 +1,77 @@
+---
+library_name: transformers
+language:
+- bn
+license: mit
+base_model: pyannote/speaker-diarization-3.1
+tags:
+- speaker-diarization
+- speaker-segmentation
+- bangla
+- bengali
+- pyannote
+- audio
+- generated_from_trainer
+datasets:
+- Sam3000/speaker-diarization-dataset-bangla
+model-index:
+- name: bangla-segment
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# bangla-segment
+
+This model is a fine-tuned version of [pyannote/speaker-diarization-3.1](https://huggingface.co/pyannote/speaker-diarization-3.1) on the Sam3000/speaker-diarization-dataset-bangla dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.4452
+- Model Preparation Time: 0.0056
+- Der: 0.1488
+- False Alarm: 0.0317
+- Missed Detection: 0.0372
+- Confusion: 0.0799
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 0.001
+- train_batch_size: 32
+- eval_batch_size: 32
+- seed: 42
+- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- num_epochs: 5
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss | Model Preparation Time | Der    | False Alarm | Missed Detection | Confusion |
+|:-------------:|:-----:|:----:|:---------------:|:----------------------:|:------:|:-----------:|:----------------:|:---------:|
+| 0.4657        | 1.0   | 170  | 0.4409          | 0.0056                 | 0.1506 | 0.0392      | 0.0198           | 0.0916    |
+| 0.4403        | 2.0   | 340  | 0.4201          | 0.0056                 | 0.1507 | 0.0328      | 0.0317           | 0.0861    |
+| 0.3691        | 3.0   | 510  | 0.4362          | 0.0056                 | 0.1485 | 0.0317      | 0.0350           | 0.0818    |
+| 0.3602        | 4.0   | 680  | 0.4437          | 0.0056                 | 0.1493 | 0.0319      | 0.0377           | 0.0797    |
+| 0.3875        | 5.0   | 850  | 0.4452          | 0.0056                 | 0.1488 | 0.0317      | 0.0372           | 0.0799    |
+
+
+### Framework versions
+
+- Transformers 4.46.3
+- Pytorch 2.4.1+cu118
+- Datasets 3.1.0
+- Tokenizers 0.20.3

config.json

@@ -0,0 +1,18 @@
+{
+  "architectures": [
+    "SegmentationModel"
+  ],
+  "chunk_duration": 10.0,
+  "max_speakers_per_chunk": 3,
+  "max_speakers_per_frame": 2,
+  "min_duration": null,
+  "model_type": "pyannet",
+  "sample_rate": 16000,
+  "torch_dtype": "float32",
+  "transformers_version": "4.46.3",
+  "warm_up": [
+    0.0,
+    0.0
+  ],
+  "weigh_by_cardinality": false
+}

config.yaml

@@ -0,0 +1,19 @@
+version: 3.1.0
+
+pipeline:
+  name: pyannote.audio.pipelines.SpeakerDiarization
+  params:
+    clustering: AgglomerativeClustering
+    embedding: pyannote/wespeaker-voxceleb-resnet34-LM
+    embedding_batch_size: 32
+    embedding_exclude_overlap: true
+    segmentation: Sam3000/OUTPUT_DIR
+    segmentation_batch_size: 32
+
+params:
+  clustering:
+    method: centroid
+    min_cluster_size: 12
+    threshold: 0.7045654963945799
+  segmentation:
+    min_duration_off: 0.0

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4b22a94c5458f976484c163bc918e3959456b587828b43360ed06f7f76bd973
+size 5899124

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37f7aeaf09d8f2875a4907bde327dc591bfae348eba35a307814adf532d735bc
+size 5304