Add model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
tags:
|
| 5 |
+
- tiny-random
|
| 6 |
+
- testing
|
| 7 |
+
- minicpmo
|
| 8 |
+
- optimum-intel
|
| 9 |
+
license: apache-2.0
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Tiny Random MiniCPM-o-2_6
|
| 13 |
+
|
| 14 |
+
This is a tiny random version of [openbmb/MiniCPM-o-2_6](https://huggingface.co/openbmb/MiniCPM-o-2_6) generated for testing purposes in [Optimum-Intel](https://github.com/huggingface/optimum-intel).
|
| 15 |
+
|
| 16 |
+
## Model Details
|
| 17 |
+
|
| 18 |
+
- **Architecture**: `MiniCPMO` (Qwen2 + SigLIP + Whisper + ChatTTS)
|
| 19 |
+
- **Hidden Size**: 16
|
| 20 |
+
- **Num Layers**: 2
|
| 21 |
+
- **Vocab Size**: 151,700 (Same as original for tokenizer compatibility)
|
| 22 |
+
- **Parameters**: ~5M
|
| 23 |
+
- **Size**: ~9.6 MB (safetensors)
|
| 24 |
+
|
| 25 |
+
## Usage
|
| 26 |
+
|
| 27 |
+
This model contains random weights and is intended **only for integration testing**, not for inference. It produces random output.
|
| 28 |
+
|
| 29 |
+
### How it was generated
|
| 30 |
+
|
| 31 |
+
Generated using a custom script that:
|
| 32 |
+
|
| 33 |
+
1. Loads the original config.
|
| 34 |
+
2. Shrinks hidden sizes to 16 and layers to 1-2.
|
| 35 |
+
3. Disables TTS initialization (due to large hardcoded layers in DVAE).
|
| 36 |
+
4. Initializes random weights.
|