Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,5 @@
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
tags:
|
|
@@ -28,72 +30,6 @@ This test model is exclusively for the Meta's LLaMA 3.2 1B (1024 context) model
|
|
| 28 |
|
| 29 |
---
|
| 30 |
|
| 31 |
-
## Requirements
|
| 32 |
-
|
| 33 |
-
- **macOS Sequoia** with Apple Neural Engine and 16GB RAM
|
| 34 |
-
- **CoreML Tools** and **HuggingFace Transformers** libraries
|
| 35 |
-
- **Python 3.9**
|
| 36 |
-
|
| 37 |
-
`chat.py` provides a sample inference script.
|
| 38 |
-
`chat_full.py` provides a sample inference script with history and conversation management.
|
| 39 |
-
|
| 40 |
-
**Installation**
|
| 41 |
-
|
| 42 |
-
1. Download the model from Hugging Face:
|
| 43 |
-
```bash
|
| 44 |
-
# Install required tools
|
| 45 |
-
pip install huggingface_hub
|
| 46 |
-
|
| 47 |
-
# Install Git LFS (Large File Support)
|
| 48 |
-
# macOS with Homebrew:
|
| 49 |
-
brew install git-lfs
|
| 50 |
-
# Or Ubuntu/Debian:
|
| 51 |
-
# sudo apt-get install git-lfs
|
| 52 |
-
|
| 53 |
-
# Initialize Git LFS
|
| 54 |
-
git lfs install
|
| 55 |
-
|
| 56 |
-
# Clone the repository with model files
|
| 57 |
-
git clone https://huggingface.co/anemll/anemll-DeepSeekR1-8B-ctx1024_0.2.0
|
| 58 |
-
```
|
| 59 |
-
|
| 60 |
-
2. Extract model files:
|
| 61 |
-
```bash
|
| 62 |
-
# Navigate to cloned directory
|
| 63 |
-
cd anemll-DeepSeekR1-8B-ctx1024_0.2.0
|
| 64 |
-
|
| 65 |
-
# Pull LFS files (model weights)
|
| 66 |
-
git lfs pull
|
| 67 |
-
|
| 68 |
-
# Extract CoreML model files
|
| 69 |
-
find . -type f -name "*.zip" -exec unzip {} \;
|
| 70 |
-
```
|
| 71 |
-
|
| 72 |
-
3. Install dependencies:
|
| 73 |
-
```bash
|
| 74 |
-
pip install coremltools transformers
|
| 75 |
-
```
|
| 76 |
-
|
| 77 |
-
**Coremltools:**
|
| 78 |
-
|
| 79 |
-
See coremltools installation guide at https://coremltools.readme.io/v4.0/docs/installation
|
| 80 |
-
|
| 81 |
-
**How to Run**
|
| 82 |
-
|
| 83 |
-
1. Basic chat interface:
|
| 84 |
-
```bash
|
| 85 |
-
python chat.py --meta ./meta.yaml
|
| 86 |
-
```
|
| 87 |
-
|
| 88 |
-
2. Full conversation mode with history:
|
| 89 |
-
```bash
|
| 90 |
-
python chat_full.py --meta ./meta.yaml
|
| 91 |
-
```
|
| 92 |
-
|
| 93 |
-
> Note: The first time the model loads, macOS will take some time to place it on the device.
|
| 94 |
-
> Subsequent loads will be instantaneous.
|
| 95 |
-
> Use Ctrl-D to exit, Ctrl-C to interrupt inference.
|
| 96 |
-
|
| 97 |
**More Info**
|
| 98 |
Please check following links for later updates:
|
| 99 |
|
|
|
|
| 1 |
+
TEST ONLY, this model requires iPad Pro > 8GB of meomry (16GB)
|
| 2 |
+
|
| 3 |
---
|
| 4 |
license: mit
|
| 5 |
tags:
|
|
|
|
| 30 |
|
| 31 |
---
|
| 32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
**More Info**
|
| 34 |
Please check following links for later updates:
|
| 35 |
|