card: full rewrite from canonical template

735abf9 verified 3 days ago

3.77 kB

license: apache-2.0
language:
  - en
  - pl
base_model:
  - tomoro-ai/Colqwen3-8B-base
library_name: mlx
pipeline_tag: visual-document-retrieval
tags:
  - mlx
  - colbert
  - visual-retrieval
  - document-understanding
  - apple-silicon
  - qwen3-vl
  - qwen
inference: false
datasets:
  - vidore/colpali_train_set

colqwen3-8b-vetcoders-mlx

colqwen3-8b-vetcoders-mlx is an MLX visual document retrieval model derived from tomoro-ai/Colqwen3-8B-base, built for image/page and text-query embedding workflows.

Intended use

Visual document retrieval over page images and text queries
Late-interaction ranking experiments for PDFs, scans, and visually rich documents
Apple Silicon local retrieval pipelines that need MLX-native weights

Out of scope

Safety-critical decisions without domain expert review
Claims of benchmark superiority not backed by published evaluation data
Non-MLX runtime guarantees; this card documents the shipped HF checkpoint, not every possible serving stack
High-stakes visual interpretation without human review

Training and conversion metadata

Parameter	Value
Repository	`LibraxisAI/colqwen3-8b-vetcoders-mlx`
Base model	`tomoro-ai/Colqwen3-8B-base`
Task	`visual-document-retrieval`
Library	`mlx`
Format	MLX / Apple Silicon checkpoint
Quantization	Not declared
Architecture	Qwen3VLForConditionalGeneration
Model files	9
Config model_type	`qwen3_vl`

This card only reports metadata present in the Hugging Face repository, existing card frontmatter, or public config files. Missing benchmark, dataset, or training-run details are left explicit rather than reconstructed.

Usage

Python

# Example shape for MLX document-retrieval workflows.
# Use the model-specific retrieval wrapper in your application code.
model_id = "LibraxisAI/colqwen3-8b-vetcoders-mlx"
query = "Which page discusses treatment protocol changes?"
document_image = "page.png"

Notes

This checkpoint is for retrieval embeddings rather than free-form chat.
Pair it with a ColBERT/MaxSim-style ranking implementation that supports the model layout.

Example output

No public sample output is currently declared for this checkpoint. Run the usage example above against your own prompt or audio/image input to inspect behavior.

Comparison with the base model

Aspect	Base	This checkpoint
Lineage	`tomoro-ai/Colqwen3-8B-base`	`LibraxisAI/colqwen3-8b-vetcoders-mlx`
Runtime target	Upstream runtime format	MLX on Apple Silicon
Published benchmark delta	Not declared in public metadata	Not declared in public metadata

Limitations

No public benchmarks for this checkpoint are declared in the model metadata.
No public benchmark claims are made by this card unless listed in the frontmatter.
Validate outputs on your own domain data before relying on this checkpoint.
Memory use and speed depend heavily on the exact Apple Silicon generation, unified-memory size, and prompt length.

License

apache-2.0. Check the upstream/base model license as well when a base model is declared.

Citation

@misc{libraxisai-colqwen3-8b-vetcoders-mlx,
  title = {colqwen3-8b-vetcoders-mlx},
  author = {LibraxisAI},
  year = {2026},
  howpublished = {\url{https://huggingface.co/LibraxisAI/colqwen3-8b-vetcoders-mlx}},
  note = {MLX checkpoint published by LibraxisAI}
}

Inference tested on

LibraxisAI/mlx-batch-server

Base model: tomoro-ai/Colqwen3-8B-base

LibraxisAI
/

colqwen3-8b-vetcoders-mlx