Update README.md
Browse files
README.md
CHANGED
|
@@ -18,19 +18,6 @@ The PlanPhys model family explores different pre-training strategies, including
|
|
| 18 |
|
| 19 |
---
|
| 20 |
|
| 21 |
-
## 🚀 Model Zoo
|
| 22 |
-
|
| 23 |
-
We provide multiple pretrained variants of PlanPhys models:
|
| 24 |
-
|
| 25 |
-
| Model | Description |
|
| 26 |
-
|---|---|
|
| 27 |
-
| 🔹 [PlanPhys-Model-Pre-Training-Random](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Random) | Baseline model pretrained with random data sampling strategy |
|
| 28 |
-
| 🔹 [PlanPhys-Model-Pre-Training-Scaling](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Scaling) | Scaled-up pretraining model with larger training resources and data |
|
| 29 |
-
| 🔹 [PlanPhys-Model-Pre-Training-Distribution](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Distribution) | Model optimized with improved data distribution strategies |
|
| 30 |
-
| 🔹 [PlanPhys-Model-Pre-Training-Ensemble](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Ensemble) | Ensemble model combining multiple pretrained representations |
|
| 31 |
-
|
| 32 |
-
---
|
| 33 |
-
|
| 34 |
## 📌 Overview
|
| 35 |
|
| 36 |
PlanPhys investigates how large-scale pretraining strategies influence the learning of physical knowledge.
|
|
@@ -53,29 +40,21 @@ The released models cover different perspectives:
|
|
| 53 |
|
| 54 |
---
|
| 55 |
|
| 56 |
-
## 🛠️ Usage
|
| 57 |
|
| 58 |
-
|
| 59 |
|
| 60 |
-
|
| 61 |
-
- Configuration files
|
| 62 |
-
- Training details
|
| 63 |
-
- Inference examples
|
| 64 |
|
| 65 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
|
| 67 |
---
|
| 68 |
|
| 69 |
-
## 📚 Related Resources
|
| 70 |
|
| 71 |
-
| Resource | Link |
|
| 72 |
-
|---|---|
|
| 73 |
-
| Random Pretraining | https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Random |
|
| 74 |
-
| Scaling Study | https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Scaling |
|
| 75 |
-
| Distribution Study | https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Distribution |
|
| 76 |
-
| Ensemble Model | https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Ensemble |
|
| 77 |
-
|
| 78 |
-
---
|
| 79 |
|
| 80 |
## 📝 Citation
|
| 81 |
|
|
|
|
| 18 |
|
| 19 |
---
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
## 📌 Overview
|
| 22 |
|
| 23 |
PlanPhys investigates how large-scale pretraining strategies influence the learning of physical knowledge.
|
|
|
|
| 40 |
|
| 41 |
---
|
| 42 |
|
|
|
|
| 43 |
|
| 44 |
+
## 🚀 Model Zoo
|
| 45 |
|
| 46 |
+
We provide multiple pretrained variants of PlanPhys models:
|
|
|
|
|
|
|
|
|
|
| 47 |
|
| 48 |
+
| Model | Description |
|
| 49 |
+
|---|---|
|
| 50 |
+
| 🔹 [PlanPhys-Model-Pre-Training-Random](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Random) | Baseline model pretrained with random data sampling strategy |
|
| 51 |
+
| 🔹 [PlanPhys-Model-Pre-Training-Scaling](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Scaling) | Scaled-up pretraining model with larger training resources and data |
|
| 52 |
+
| 🔹 [PlanPhys-Model-Pre-Training-Distribution](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Distribution) | Model optimized with improved data distribution strategies |
|
| 53 |
+
| 🔹 [PlanPhys-Model-Pre-Training-Ensemble](https://huggingface.co/MultimodalAgent/PlanPhys-Model-Pre-Training-Ensemble) | Ensemble model combining multiple pretrained representations |
|
| 54 |
|
| 55 |
---
|
| 56 |
|
|
|
|
| 57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
## 📝 Citation
|
| 60 |
|