vrfai/molmoact2-libero-gguf

robotics model by vrfai

Older snapshot: check the original source for current information.

From the publisher

Model card & documentation

Source preview

Read the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.

This is a GGUF conversion of allenai/MolmoAct2-LIBERO (paper, code). It runs on vla.cpp, a lightweight C++ inference engine for Vision-Language-Action models built on top of llama.cpp. A SigLIP-so400m/14 tower encodes each camera at 378×378; the outputs of its layers 24 and 18 are pooled 2×2 by attention into 196 tokens per view and added onto the image-patch embeddings of a 36-layer Qwen3-style LM (2560 wide, 32 q / 8 kv heads). A flow-matching action expert with one adaLN block per LM layer cross-attends to that layer's keys and values and integrates a 10×7 action chunk in 10 Euler steps. The robot state enters as discrete tokens in the prompt. The tower, the LM and the expert are all in the single GGUF, so there is no mmproj file. Only the tokenizer runs outside the engine. One policy fine-tuned on all four LIBERO suites (spatial, object, goal, 10). Continuous actions only:…

vrfai/molmoact2-libero-gguf on Hugging Face A short preview, not the full document.

Read the full model card ↗ · Preview checked 2026-11-10T09:16:36.791Z

Inside the original model card — Document outline

Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.

Links from the model card

References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.

paper

Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.

Model overview

vrfai/molmoact2-libero-gguf is a robotics model repository published by vrfai on Hugging Face. The source reports the vla.cpp library.

This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.

Model facts

Task: robotics
Library: vla.cpp
Recent downloads (30 days): 0
Cumulative likes: 0
Hugging Face trending score: 0
Reported safetensors parameters: Not reported
Architecture: MolmoAct2ForConditionalGeneration
License: apache-2.0
Access: Not gated by Hugging Face
Created: 2026-11-10T06:52:33.000Z
Last modified: 2026-11-10T06:53:49.000Z

Compatibility and lineage

Language coverage was not reported.

Reported base models: allenai/MolmoAct2-LIBERO

Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.

Use and evaluate

Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.

No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.

Model card and usage instructions

Model files and configuration

Community discussion