Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
base_model:
|
| 3 |
-
-
|
| 4 |
tags:
|
| 5 |
- transformers
|
| 6 |
- unsloth
|
|
@@ -29,7 +29,7 @@ library_name: transformers
|
|
| 29 |
<br/><br/>
|
| 30 |
|
| 31 |
<a href="https://colab.research.google.com/drive/17ZeVLF8Q1P9juIKftijN9uwl2LrfMeCz?usp=drive_link">
|
| 32 |
-
📓 Colab Notebook: Orpheus 3B TTS Finetuning
|
| 33 |
</a>
|
| 34 |
</div>
|
| 35 |
|
|
@@ -40,16 +40,32 @@ This repository contains a fine-tuned variant of **Orpheus TTS 3B** — a state-
|
|
| 40 |
|
| 41 |
# **Fine-Tuned Model Description**
|
| 42 |
|
| 43 |
-
This model is derived from the base checkpoint **`
|
| 44 |
|
| 45 |
Additional alignment and adaptation were performed with the **Hugging Face TRL** (Training Reinforcement Learning) library.
|
| 46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
### **Fine-Tuned Model Information**
|
| 48 |
- **Fine-tuned by:** `InnovationM's AI Team`
|
| 49 |
- **Organization:** [InnovationM](https://www.innovationm.com/)
|
| 50 |
- **License:** Apache-2.0
|
| 51 |
-
- **Base Model:** `
|
| 52 |
-
- **Dataset Used:** `aiAtInnovationM/dailytalk-male`
|
| 53 |
|
| 54 |
# **Special Thanks**
|
| 55 |
|
|
@@ -73,16 +89,15 @@ The **InnovationM** team.
|
|
| 73 |
| InnovationM LinkedIn | https://in.linkedin.com/company/innovationm |
|
| 74 |
| Fine-Tuning Notebook | https://colab.research.google.com/drive/17ZeVLF8Q1P9juIKftijN9uwl2LrfMeCz?usp=drive_link |
|
| 75 |
| Base Model | https://huggingface.co/unsloth/csm-1b |
|
| 76 |
-
| Unsloth Framework | https://github.com/unslothai/unsloth |
|
| 77 |
|
| 78 |
# 📄 **Citation**
|
| 79 |
|
| 80 |
If you use this model in your research, please cite both **Canopy Labs Model** and this **InnovationM fine-tuned version**.
|
| 81 |
|
| 82 |
```bibtex
|
| 83 |
-
@misc{
|
| 84 |
title = {Orpheus-3B Fine-Tuned Model},
|
| 85 |
-
author = {InnovationM
|
| 86 |
howpublished = {Hugging Face},
|
| 87 |
year = {2025}
|
| 88 |
}
|
|
|
|
| 1 |
---
|
| 2 |
base_model:
|
| 3 |
+
- canopylabs/orpheus-3b-0.1-ft
|
| 4 |
tags:
|
| 5 |
- transformers
|
| 6 |
- unsloth
|
|
|
|
| 29 |
<br/><br/>
|
| 30 |
|
| 31 |
<a href="https://colab.research.google.com/drive/17ZeVLF8Q1P9juIKftijN9uwl2LrfMeCz?usp=drive_link">
|
| 32 |
+
📓 Colab Notebook: Orpheus 3B TTS Finetuning
|
| 33 |
</a>
|
| 34 |
</div>
|
| 35 |
|
|
|
|
| 40 |
|
| 41 |
# **Fine-Tuned Model Description**
|
| 42 |
|
| 43 |
+
This model is derived from the base checkpoint **`canopylabs/orpheus-3b-0.1-ft`**, fine-tuned using the **Unsloth optimization framework**, which enables up to **2× faster training** with reduced memory footprint.
|
| 44 |
|
| 45 |
Additional alignment and adaptation were performed with the **Hugging Face TRL** (Training Reinforcement Learning) library.
|
| 46 |
|
| 47 |
+
### **Audio Samples**
|
| 48 |
+
|
| 49 |
+
#### **Before Finetuning (Female Voice)**
|
| 50 |
+
|
| 51 |
+
<audio controls>
|
| 52 |
+
<source src="samples/before.wav" type="audio/wav">
|
| 53 |
+
Your browser does not support the audio element.
|
| 54 |
+
</audio>
|
| 55 |
+
|
| 56 |
+
#### **After Finetuning (Male Voice)**
|
| 57 |
+
|
| 58 |
+
<audio controls>
|
| 59 |
+
<source src="samples/after.wav" type="audio/wav">
|
| 60 |
+
Your browser does not support the audio element.
|
| 61 |
+
</audio>
|
| 62 |
+
|
| 63 |
### **Fine-Tuned Model Information**
|
| 64 |
- **Fine-tuned by:** `InnovationM's AI Team`
|
| 65 |
- **Organization:** [InnovationM](https://www.innovationm.com/)
|
| 66 |
- **License:** Apache-2.0
|
| 67 |
+
- **Base Model:** `canopylabs/orpheus-3b-0.1-ft`
|
| 68 |
+
- **Dataset Used:** `aiAtInnovationM/dailytalk-male`
|
| 69 |
|
| 70 |
# **Special Thanks**
|
| 71 |
|
|
|
|
| 89 |
| InnovationM LinkedIn | https://in.linkedin.com/company/innovationm |
|
| 90 |
| Fine-Tuning Notebook | https://colab.research.google.com/drive/17ZeVLF8Q1P9juIKftijN9uwl2LrfMeCz?usp=drive_link |
|
| 91 |
| Base Model | https://huggingface.co/unsloth/csm-1b |
|
|
|
|
| 92 |
|
| 93 |
# 📄 **Citation**
|
| 94 |
|
| 95 |
If you use this model in your research, please cite both **Canopy Labs Model** and this **InnovationM fine-tuned version**.
|
| 96 |
|
| 97 |
```bibtex
|
| 98 |
+
@misc{innovationm2025,
|
| 99 |
title = {Orpheus-3B Fine-Tuned Model},
|
| 100 |
+
author = {InnovationM},
|
| 101 |
howpublished = {Hugging Face},
|
| 102 |
year = {2025}
|
| 103 |
}
|