hrithik-dev8/tiny-random-MiniCPM-o-2_6

Tiny Random MiniCPM-o-2_6

A tiny (~42 MB) randomly-initialized version of MiniCPM-o-2.6 designed for testing purposes in the optimum-intel library.

Purpose

This model was created to replace the existing test model at optimum-intel-internal-testing/tiny-random-MiniCPM-o-2_6 (185 MB) with a smaller alternative for CI/CD testing. Smaller test models reduce:

• Download times in CI pipelines
• Storage requirements
• Test execution time

Size Comparison

Model	Total Size	Model Weights
openbmb/MiniCPM-o-2_6 (Original)	17.4 GB	~17 GB
optimum-intel-internal-testing/tiny-random-MiniCPM-o-2_6 (Current Test Model)	185 MB	169 MB
hrithik-dev8/tiny-random-MiniCPM-o-2_6 (This Model)	~42 MB	41.55 MB

Result: 4× smaller than Intel's current test model

Model Configuration

Component	This Model	Original
Vocabulary	5,000 tokens	151,700 tokens
LLM Hidden Size	128	3,584
LLM Layers	1	40
LLM Attention Heads	8	28
Vision Hidden Size	128	1,152
Vision Layers	1	27
Image Size	980 (preserved)	980
Patch Size	14 (preserved)	14
Audio d_model	64	1,280
TTS Hidden Size	128	-

Parameter Breakdown

Component	Parameters	Size (MB)
TTS/DVAE	19,339,766	36.89
LLM	1,419,840	2.71
Vision	835,328	1.59
Resampler	91,392	0.17
Audio	56,192	0.11
Other	20,736	0.04
Total	21,763,254	~41.5

Technical Details

Why Keep TTS/DVAE Components?

The TTS (Text-to-Speech) component, which includes the DVAE (Discrete Variational Auto-Encoder), accounts for approximately 37 MB (~85%) of the model size. While the optimum-intel tests do not exercise TTS functionality (they only test image+text → text generation), we retain this component because:

1. Structural Consistency: Removing TTS via init_tts=False causes structural differences in the model that lead to numerical divergence between PyTorch and OpenVINO outputs
2. Test Compatibility: The test_compare_to_transformers test compares PyTorch vs OpenVINO outputs and requires exact structural matching
3. Architecture Integrity: The MiniCPM-o architecture expects TTS weights to be present during model loading

Tokenizer Shrinking

The vocabulary was reduced from 151,700 to 5,000 tokens:

• Base tokens: IDs 0-4899 (first 4,900 most common tokens)
• Special tokens: IDs 4900-4949 (remapped from original high IDs)
• BPE merges: Filtered from 151,387 to 4,644 (only merges involving retained tokens)

Key token mappings:

Token	ID
<unk>	4900
<|endoftext|>	4901
<|im_start|>	4902
<|im_end|>	4903

Reproducibility

Model weights are initialized with a fixed random seed (42) to ensure:

• Reproducible outputs between runs
• Consistent behavior between PyTorch and OpenVINO
• Passing of test_compare_to_transformers which compares framework outputs

Test Results

Tested with pytest tests/openvino/test_seq2seq.py -k "minicpmo" -v:

Test	Status	Notes
test_compare_to_transformers	✅ PASSED	PyTorch/OpenVINO outputs match
test_generate_utils	✅ PASSED	Generation pipeline works
test_model_can_be_loaded_after_saving	⚠️ FAILED	Windows file locking issue (not model-related)

The third test failure is a Windows-specific issue where OpenVINO keeps file handles open, preventing cleanup of temporary directories. This is a known platform limitation, not a model defect. The test passes on Linux/macOS.

Usage

For optimum-intel Testing

# In optimum-intel/tests/openvino/utils_tests.py, update MODEL_NAMES:
MODEL_NAMES = {
    # ... other models ...
    "minicpmo": "hrithik-dev8/tiny-random-MiniCPM-o-2_6",
}

Then run tests:

pytest tests/openvino/test_seq2seq.py -k "minicpmo" -v

Basic Model Loading

from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained(
    "hrithik-dev8/tiny-random-MiniCPM-o-2_6",
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained(
    "hrithik-dev8/tiny-random-MiniCPM-o-2_6",
    trust_remote_code=True
)

Files Included

File	Size	Description
model.safetensors	41.55 MB	Model weights (bfloat16)
config.json	5.33 KB	Model configuration
tokenizer.json	338.27 KB	Shrunk tokenizer (5,000 tokens)
tokenizer_config.json	12.78 KB	Tokenizer settings
vocab.json	85.70 KB	Vocabulary mapping
merges.txt	36.58 KB	BPE merge rules
preprocessor_config.json	1.07 KB	Image processor config
generation_config.json	121 B	Generation settings
added_tokens.json	1.13 KB	Special tokens
special_tokens_map.json	1.24 KB	Special token mappings

Requirements

• Python 3.8+
• transformers >= 4.45.0, < 4.52.0
• torch
• For OpenVINO testing: optimum-intel with OpenVINO backend

Limitations

⚠️ This model is for testing only - it produces random/meaningless outputs and should not be used for inference.