Update README.md
Browse files
README.md
CHANGED
|
@@ -30,7 +30,7 @@ learning from human feedback (RLHF).
|
|
| 30 |
---
|
| 31 |
|
| 32 |
## 🧩 Tasks
|
| 33 |
-
|
| 34 |
impressions (e.g., the mood of a scene), and associative indicates real-world familiar experiences (e.g., buzzing of a bee, a heartbeat), the goal is to generate a caption corresponding to the specified category of haptic experience.
|
| 35 |
|
| 36 |
---
|
|
@@ -42,9 +42,9 @@ HapticLLaMA training is consist of (1) supervised fine-tuning with LoRA adaptati
|
|
| 42 |
|
| 43 |
|
| 44 |
- ## 📂 Models
|
| 45 |
-
- **Frequency-based Model
|
| 46 |
|
| 47 |
-
- **Encodec-based Model
|
| 48 |
---
|
| 49 |
|
| 50 |
## 📂 Haptic Tokenizer
|
|
|
|
| 30 |
---
|
| 31 |
|
| 32 |
## 🧩 Tasks
|
| 33 |
+
Given a vibration signal S and a target category c ∈ {sensory, emotional, associative}, where sensory refers to physical attributes (e.g.,intensity of tapping), emotional denotes affective
|
| 34 |
impressions (e.g., the mood of a scene), and associative indicates real-world familiar experiences (e.g., buzzing of a bee, a heartbeat), the goal is to generate a caption corresponding to the specified category of haptic experience.
|
| 35 |
|
| 36 |
---
|
|
|
|
| 42 |
|
| 43 |
|
| 44 |
- ## 📂 Models
|
| 45 |
+
- **Frequency-based Model**
|
| 46 |
|
| 47 |
+
- **Encodec-based Model**
|
| 48 |
---
|
| 49 |
|
| 50 |
## 📂 Haptic Tokenizer
|