Downloads · 30 days
45
7% of all-time downloads
ysdede/Phi-4-mm-inst-asr-turkish-unf
Phi-4-mm-inst-asr-turkish-unf is a text generation model from ysdede. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
Downloads · 30 days
45
7% of all-time downloads
All-time downloads
684
Public
Parameters
5.6B
11.1 GB on disk
Likes
4
Public
Click a slice to open those files.
.safetensors11.1 GB · 100%
From the Hugging Face model README
This model is a fine-tuned version of microsoft/Phi-4-multimodal-instruct.
Model Background:
This benchmark evaluates a fine-tuned version of Microsoft's Phi-4-mm-instruct, a multimodal model not originally designed for Turkish ASR. Key points:
Initial Limitations:
Fine-Tuning Process:
Current Status:
Why This Matters:
Testing Environment: Google Colab with L4 GPU (24 GB VRAM)
| Model | WER (%) | CER (%) | Inference Speed (xRT) | Batch Size | Audio Duration (hrs) | Samples Processed |
|---|---|---|---|---|---|---|
| ysdede/commonvoice_17_tr_fixed | 9.7 | 2.72 | x26 | 32 | 7.1 | 8,576 |
| erenfazlioglu/turkishvoicedataset | 11.52 | 3.93 | x20 | 16 | 7.8 | 2,496 |
| ysdede/khanacademy-turkish | 12.04 | 7.78 | x16 | 16 | 3.8 | 1,344 |
| ysdede/yeni-split-0 | 20.58 | 13.2 | x16 | 16 | 18 | 5,936 |
| ymoslem/MediaSpeech | 25.48 | 15.16 | x35 | 32 | 10 | 2,496 |
| dssnt1 | 27.23 | 9.6 | x12 | 16 | 2.5 | 1,200 |
| ysdede/yeni-split-lq-noisy | 39.4 | 27 | x19 | 16 | 12 | 3,440 |
Dataset Notes:
Finetuning Datasets:
commonvoice_17_tr_fixed: Crowd-sourced clean speech (not professional studio recordings) with shuffled splits - potential speaker leakage (same speakers in train/test with different utterances)khanacademy-turkish: Educational lectures with STEM vocabularyyeni-split-0: Noisy real-world recordingsBenchmark-only Datasets:
turkishvoicedataset: Synthetic TTS news (clean but artificial prosody)yeni-split-lq-noisy: Challenging noisy samples with alignment errorsText Normalization Challenges:
⚠️ Current WER/CER scores may be inflated due to:
Evaluation Note:
For Turkish ASR benchmarking, I developed a text normalizer to address language-specific scoring challenges. While imperfect, it helps:
This preprocessing makes WER/CER calculations slightly fairer compared to raw scoring, though manual verification remains recommended. The tool is actively being refined based on validation set findings.
Performance Factors:
More information needed
More information needed
More information needed
The following hyperparameters were used during training: