JMNI-Labs/Inkling-2-Mini-FLARE
image-text-to-text model by JMNI-Labs
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
Prepared Flare package that serves thinkingmachines/Inkling-2-Mini (24 physical → 48 logical looped MoE layers, NVFP4 layers 0–19 + MXFP8 layers 20–23 SiTU experts, vision, TML tool use, and bundled 3-layer DSpark 7-draft speculative decoding) on one DGX Spark (NVIDIA GB10, SM121, 121 GiB unified memory). main holds this model card, the Apache-2.0 license, publisher.json, and the b12x patch under patches/. The ready-to-run core lives on branch gb10-tp1. Served with --no-history-drafts --mtp-adaptive-verification --mtp-adaptive-verification-alpha 0.5. Temperature 1, 20 s cells, max 2048 output tokens, --coding-peak, --no-hw-monitor: Aggregate decode throughput (tok/s): 8 fixed prompts (code, code2, prose, explain, math, JSON, essay, SQL), 512 output tokens, thinking off (tok/s): tests/server/testinkling2.py passed 14/14 against this exported core on native flare run (both default reasoning effort and zero reasoning effort): single and 8-way…
Read the full model card ↗ · Preview checked 2026-11-21T08:14:41.160Z
Inside the original model card — Document outline
- Inkling 2 Mini on a single DGX Spark — FLARE TP1 with 7-Draft DSpark
- Results (single ASUS GX10 / DGX Spark, 0–2350 MHz GPU ceiling)
- Local Inference Lab llm-inference-bench 0.8.0
- Fixed-seed A/B probe (abbench.py)
- Qualification
- Running this package
- Build provenance & patches
- License
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.
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
JMNI-Labs/Inkling-2-Mini-FLARE is a image-text-to-text model repository published by JMNI-Labs on Hugging Face. A library was not reported.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- image-text-to-text
- Library
- Not reported
- Recent downloads (30 days)
- 0
- Cumulative likes
- 0
- Hugging Face trending score
- 0
- Reported safetensors parameters
- Not reported
- Architecture
- Not reported
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-11-21T06:55:22.000Z
- Last modified
- 2026-11-21T06:55:22.000Z
Compatibility and lineage
Language coverage was not reported.
Reported base models: thinkingmachines/Inkling-2-Mini
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.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-11-21T07:02:33.996Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.