Technical report: Read paper · PDF · BibTeX · All models
Collection: Mermaid models
Usage: Step-by-step Windows, Linux and macOS instructions.
MiniCPM-V 4.7 1B Mermaid — LoRA Adapter
GGUF: Lucasgitguy-co/minicpm-v-4-7-1b-mermaid-adapter-GGUF (LoRA for llama.cpp).
LoRA adapter repository for the MiniCPM-V 4.7 Mermaid project. This is an adapter, not a standalone set of base-model weights.
| File | Size (decimal GB) |
|---|---|
adapter_model.safetensors |
0.020 |
Adapter configuration
- LoRA rank: 8
- LoRA alpha: 16
- LoRA dropout: 0.05
- PEFT version recorded in configuration: 0.21.2
- Base model class:
MiniCPMV4_7ForConditionalGeneration
The adapter configuration resolves the original base checkpoint as openbmb/MiniCPM-V-4.7-1B, pinned to revision 286817f1132b998913a129a221c89b4402b4ab80. This revision was verified against the local base download metadata and matching configuration. Initialize that base model and its processor before applying this adapter.
Download and loading pattern
pip install -U huggingface_hub
hf download Lucasgitguy-co/minicpm-v-4-7-1b-mermaid-adapter --local-dir ./minicpm-v-4-7-1b-mermaid-adapter
After initializing base_model with the original compatible model implementation and weights:
from peft import PeftModel
model = PeftModel.from_pretrained(
base_model,
"Lucasgitguy-co/minicpm-v-4-7-1b-mermaid-adapter",
)
model.eval()
Use the original base model’s image processor and inference interface. The base model must exist before this snippet is run.
Evaluation and provenance
This card describes the repository files and configuration. Detailed training data, training procedure, and benchmark results are not documented in the inspected artifacts. Loading examples are guidance, not a claim of measured model quality.
License and attribution
This release is licensed under Apache 2.0; see LICENSE and NOTICE. Upstream attribution and applicable third-party notices must be retained when redistributing.
Repository guide
Related project data: text and image collection and original English/Dutch text collection. These links describe related project resources; they do not establish the exact training split of every checkpoint.