File size: 2,762 Bytes
292d92c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 |
---
description: Set up conda environment for speech-to-text fine-tuning
tags: [python, conda, stt, whisper, speech, ai, fine-tuning, project, gitignored]
---
You are helping the user set up a conda environment for speech-to-text (STT) fine-tuning.
## Process
1. **Create base environment**
```bash
conda create -n stt-finetune python=3.11 -y
conda activate stt-finetune
```
2. **Install PyTorch with ROCm**
```bash
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.0
```
3. **Install Whisper and related libraries**
```bash
pip install openai-whisper
pip install faster-whisper # Optimized inference
pip install whisperx # Advanced features
```
4. **Install Hugging Face libraries**
```bash
pip install transformers
pip install datasets
pip install accelerate
pip install evaluate
pip install peft # For LoRA fine-tuning
```
5. **Install audio processing libraries**
```bash
pip install librosa # Audio analysis
pip install soundfile # Audio I/O
pip install pydub # Audio manipulation
pip install sox # Audio processing
conda install -c conda-forge ffmpeg -y # Audio conversion
```
6. **Install speech-specific tools**
```bash
pip install jiwer # Word Error Rate calculation
pip install speechbrain # Speech toolkit
pip install pyannote.audio # Speaker diarization
```
7. **Install data processing tools**
```bash
pip install pandas
pip install numpy
pip install scipy
pip install matplotlib
pip install seaborn # Visualization
```
8. **Install monitoring and experimentation**
```bash
pip install wandb # Experiment tracking
pip install tensorboard
```
9. **Install Jupyter for interactive work**
```bash
conda install -c conda-forge jupyter jupyterlab ipywidgets -y
```
10. **Test installation**
```python
import torch
import whisper
import librosa
from transformers import WhisperProcessor, WhisperForConditionalGeneration
print(f"PyTorch: {torch.__version__}")
print(f"GPU available: {torch.cuda.is_available()}")
print("All libraries imported successfully!")
```
11. **Suggest common datasets**
- Common Voice (Mozilla)
- LibriSpeech
- TEDLIUM
- Custom datasets
12. **Create example script**
- Offer to create `~/scripts/whisper-finetune-example.py` with basic setup
## Output
Provide a summary showing:
- Environment name and setup status
- Installed libraries grouped by purpose
- GPU detection status
- Available VRAM for training
- Suggested datasets for fine-tuning
- Example commands for testing
- Links to documentation/tutorials
|