Technical report: Read paper · PDF · BibTeX · All models
Collection: Mermaid models
Usage: Step-by-step Windows, Linux and macOS instructions.
MiniMind-3 Mermaid — Run 2 Epoch Checkpoints
GGUF: Lucasgitguy-co/minimind-mermaid-run2-GGUF (of ep3 for llama.cpp).
MiniMind Mermaid run 2 checkpoint collection. The repository stores separate ep1, ep2, and ep3 folders, each with model weights, configuration, and tokenizer files.
| File | Size (decimal GB) |
|---|---|
ep1/model.safetensors |
0.256 |
ep2/model.safetensors |
0.256 |
ep3/model.safetensors |
0.256 |
Load a checkpoint
The inspected ep3/config.json declares Qwen3ForCausalLM, float16 weights, 8 hidden layers, hidden size 768, and vocabulary size 6,400. This export is a text-generation model; no vision projector is included.
from transformers import AutoTokenizer, AutoModelForCausalLM
repo = "Lucasgitguy-co/minimind-mermaid-run2"
checkpoint = "ep3" # Select ep1, ep2, or ep3 explicitly.
tokenizer = AutoTokenizer.from_pretrained(repo, subfolder=checkpoint)
model = AutoModelForCausalLM.from_pretrained(
repo, subfolder=checkpoint, torch_dtype="auto"
)
The export records Transformers 4.57.6. ep3 is an example selection, not a documented best checkpoint.
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.