← back to catalog · registered 2026-11-16 00:03

balllvin/LFM3-Omni-2B-Abliterated-Arete-ONNX-WebGPU

balllvin 2B multimodal second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/balllvin%2FLFM3-Omni-2B-Abliterated-Arete-ONNX-WebGPU"
Response includes
  • classification m1
  • files 11
  • author_summary 28 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · 30-day
13
Likes
0
Model age
6d ago
created 2026-11-15
Downloads over time
Now 13 → from 0 ↑ 0%
051014 0 on Nov 16 13 on Nov 22 13 on Nov 18 Nov
Nov 16 → Nov 22 · 7 snapshots · spans 6 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
other
Tags
onnxruntime onnx lfm3_omni lfm3 webgpu onnxruntime-web image-text-to-text conversational custom_code base_model:balllvin/LFM3-Omni-2B-Abliterated-Arete-NVFP4-MLX base_model:quantized:balllvin/LFM3-Omni-2B-Abliterated-Arete-NVFP4-MLX license:other

Related

Total size
0 B
Files
11
Quantizations
1
Registered
2026-11-16 00:03
Last updated on HF
2026-11-16 12:33

Files by quantization

Auxiliary files 11 files 17.1 MB
tokenizer.json 17.1 MB ******** download
chat_template.jinja 13.2 KB 7601ed47 download
LICENSE 10.3 KB 25e731c6 download
config.json 5.77 KB f135feb4 download
README.md 3.24 KB 51bcfae9 download
processor_config.json 2.32 KB 13e75441 download
NOTICE.md 1.75 KB 96865ca3 download
.gitattributes 1.66 KB 57c8e9e9 download
tokenizer_config.json 674 B a6b5b1dc download
preprocessor_config.json 536 B 4f6fa7e3 download
generation_config.json 392 B 9db0b0d3 download

README current version from Hugging Face


library_name: onnxruntime
license: other
license_name: lfm1.0
license_link: LICENSE
pipeline_tag: image-text-to-text
base_model: balllvin/LFM3-Omni-2B-Abliterated-Arete-NVFP4-MLX
base_model_relation: quantized
tags:

  • lfm3
  • onnx
  • webgpu
  • onnxruntime-web

LFM3-Omni-2B-Abliterated-Arete-ONNX-WebGPU

balllvin/LFM3-Omni-2B-Abliterated-Arete-NVFP4-MLX as ONNX for the browser (WebGPU,
onnxruntime-web). Chat with it in the browser: LFM3-WebGPU.
It answers without a thinking block, as the source release does. Text and images in, text out.

Parity with the bf16 source

Check Result
Same top token as bf16, all positions 94.5%
Same top token, generated part 97.5%
Mean KL from bf16 (nats per token) 0.0327
Greedy tokens equal to bf16 before the first difference 35 of 209
Reference for the same check on LiquidAI's own q4f16 vs its bf16 93.1% top token, KL 0.0503

Six fixed chat prompts, 48 greedy tokens each from transformers bf16, teacher forced through the ONNX graph with
the caches carried across prefill chunks of 32 tokens (ONNX Runtime CPU).

Speed

256 new tokens, the same prompt, median of 3 runs.

Where Decode tok/s Prefill tok/s
Apple M5 Pro, headless Chrome 15.0 227

How it was built

  • Source weights: the bf16 task vector graft rebuilt with brainspa graft build --method task_vector (host LiquidAI/LFM3-Omni-2B, donor balllvin/lfm3-1.7b-arete-ckpt ar1/best, base LiquidAI/LFM3-1.7B); its sha256 230c9167baead0e521450939961bef070f7e9a1f64ceaac0e5d9f530c29d4bb3 equals the model.safetensors sha256 in the release's graft.json.
  • Graph: LiquidAI's ONNX Runtime GenAI graph, LiquidAI/LFM3-Omni-2B-ONNX@d2e656f0bddfaa2b8ebd4bfbe9d88de7897ee592 onnx/decoder_q4f16.onnx, unchanged. Only the weights differ.
  • Quantization: the same recipe as that graph: ONNX Runtime k-quant, asymmetric, 32 weights per scale, weights
    cast to fp16 first; the embedding, the LM head and the same projections as LiquidAI's file stay in int8, the rest
    are int4. Run on LiquidAI's own bf16 source, this code gives back 379 of 594
    initializers byte for byte; the other 215 differ in a few bytes, 0.0032% of all bytes (float rounding in the scale search).
  • Vision encoder: LiquidAI/LFM3-Omni-2B-ONNX@d2e656f0bddfaa2b8ebd4bfbe9d88de7897ee592 onnx/vision_encoder_q8.onnx, byte for byte. The Omni grafts change only the language model, so vision weights are the host model's.
  • Parity is measured on the language core alone (brainspa graft extract), which is what the decoder runs for text.
  • Speed was measured in headless Chrome while the Mac was busy with other jobs (load average 10 to 12), so it is a low bound.

Files

File Size
onnx/decoder_q4f16.onnx 0.3 MB
onnx/decoder_q4f16.onnx_data 1.41 GB
onnx/vision_encoder_q8.onnx 0.1 MB
onnx/vision_encoder_q8.onnx_data 0.20 GB

Total download 1.61 GB. Use import * as ort from 'onnxruntime-web/webgpu' and prefill in chunks of at most
1024 tokens. Inputs and outputs are those of the LiquidAI graph.

Licence

LFM Open License v1.0 (see LICENSE) and NOTICE.md, as the source release.

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration