# Jev-Style-Qwen3.5-2B-Decision-v2 by Chaoliang Yan Source: https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2 Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Jev-Style-Qwen3.5-2B-Decision-v2 (1.9B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 3.8 GB | 4.5 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 1.9 GB | 2.3 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 0.9 GB | 1.1 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 7, 2026. [Jev-Style-Qwen3.5-2B-Decision-v2 on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2/gpus) ## Model Card By Chaoliang Yan, published under apache-2.0, revision 6b84dda83feb. ### Jev-Style-Qwen3.5-2B-Decision v2 (HF BF16) Website: [jevstyle.com](https://jevstyle.com/#v2) — all JevStyle decision models, benchmarks and quickstart in one place. A Jev-style decision model for classification, routing and typed choices. Give it a state, a question and a list of options; one prefill returns a selected option with calibrated probabilities. | Build | Weight size | Inference | | --- | --- | --- | | HF BF16 · this repository | 3.76 GB | Transformers + decision client | | [GGUF](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-gguf) | 1.27–3.78 GB | Q4_K_M / Q8_0 / BF16 · llama.cpp | | [MLX BF16](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-mlx-bf16) | 3.76 GB | Apple Silicon + native MLX client | Download this build: [model.safetensors](https://huggingface.co/chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2/resolve/main/model.safetensors?download=true). The repository also includes its calibration, inference client and evaluation records. ### Results 81.20% macro accuracy on the fixed English reference panel, compared with 76.68% for v1 and 75.09% for English Laya. The results below use the CUDA reference structure: 11 real-label task groups, 3,277 decisions, equal task weights, and the same 3,100-record calibration split. Results for the released deployment formats appear further below. [Read the full model card (1,345 words)](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2/card) ## Configuration Architecture Qwen3_5ForCausalLM Context length (tokens) 262,144 Layers 24 Hidden size 2,048 Feed-forward size 6,144 Attention heads 8 Key/value heads 2 Head dimension 256 Vocabulary size 248,320 Model type qwen3_5_text ## Identity and Version Repository chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2 Publisher Chaoliang Yan Task Text generation Modality Text Library transformers Parameters 1.9B parameters Languages en Revision 6b84dda83feb79d7ed14c4c312fcb918be76cbff First published 2026-09-23 Last updated 2026-09-27 ## Files and Weights 35 files, 3.8 GB in total. The weights are 1 file totalling 3.8 GB in safetensors. Weights1 file · 3.8 GB Configuration20 files · 82.2 KB Tokenizer2 files · 20.0 MB Documentation2 files · 22.6 KB Other9 files · 561.5 KB Repository1 file · 1.7 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model.safetensors | Weights | 3.8 GB | dc0290f93034 | | SHA256SUMS.json | Configuration | 4.4 KB | — | | calibration.json | Configuration | 358 B | — | | calibration.mlx.json | Configuration | 362 B | — | | config.json | Configuration | 1.8 KB | — | | decision_config.json | Configuration | 466 B | — | | evaluation/baseline_sensitivity.json | Configuration | 2.0 KB | — | | evaluation/chart_data.json | Configuration | 4.6 KB | — | | evaluation/data_manifest.json | Configuration | 10.7 KB | — | | evaluation/deployment.json | Configuration | 1.4 KB | — | | evaluation/laya_source.json | Configuration | 97 B | — | | evaluation/laya_typed_source.json | Configuration | 113 B | — | | evaluation/reference_comparison.json | Configuration | 27.0 KB | — | | evaluation/test.metrics.json | Configuration | 4.0 KB | — | | generation_config.json | Configuration | 117 B | — | | jev_h100/__init__.py | Configuration | 85 B | — | | jev_h100/common.py | Configuration | 9.2 KB | — | | jev_h100/decide.py | Configuration | 1.0 KB | — | | jev_h100/model.py | Configuration | 8.8 KB | — | | jev_mlx_client.py | Configuration | 5.4 KB | — | | training_config.yaml | Configuration | 322 B | — | | LICENSE | Documentation | 11.3 KB | — | | README.md | Documentation | 11.2 KB | — | | chat_template.jinja | Other | 29 B | — | | figures/benchmark.png | Other | 188.9 KB | 5a3a782cf203 | | figures/benchmark.svg | Other | 22.6 KB | — | | figures/calibration.png | Other | 183.8 KB | 2a4e163354b0 | | figures/calibration.svg | Other | 29.0 KB | — | | figures/robustness.png | Other | 125.3 KB | 41c317f25853 | | figures/robustness.svg | Other | 11.8 KB | — | | requirements-mlx.txt | Other | 48 B | — | | requirements.txt | Other | 93 B | — | | .gitattributes | Repository | 1.7 KB | — | | tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 | | tokenizer_config.json | Tokenizer | 1.2 KB | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 3.8 GB [Download from Chaoliang Yan](https://huggingface.co/chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2) Released by Chaoliang Yan through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Derived from Qwen/Qwen3.5-2B-Base ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 3.8 GB | | 16-bit | 3.8 GB | | 8-bit | 1.9 GB | | 4-bit | 0.9 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Built on This Model - Quantized from[Jev-Style-Qwen3.5-2B-Decision-v2-GGUF](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-gguf) - Derived from[Jev-Style-Qwen3.5-2B-Decision-v2-GGUF](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-gguf) - Derived from[Jev-Style-Qwen3.5-2B-Decision-v2-MLX-bf16](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-mlx-bf16) ## Questions About Jev-Style-Qwen3.5-2B-Decision-v2 ### How much GPU memory does Jev-Style-Qwen3.5-2B-Decision-v2 need? About 4.5 GB at 16-bit and 1.1 GB at 4-bit: the weights (1.9B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Jev-Style-Qwen3.5-2B-Decision-v2 on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use Jev-Style-Qwen3.5-2B-Decision-v2 commercially? Yes. Jev-Style-Qwen3.5-2B-Decision-v2 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors. ### What is Jev-Style-Qwen3.5-2B-Decision-v2's context length? 262,144 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Text generation ### [Jev-Style-Qwen3.5-2B-Decision-MLX-bf16](https://savrn.com/models/jev-style-qwen3-5-2b-decision-mlx-bf16) [Chaoliang Yan](https://savrn.com/model-publishers/chaoliangunsw) Website: jevstyle.com — all JevStyle decision models, benchmarks and quickstart in one place. A Jev-style decision model: it does not write text. Give it a state, a question and a list of options, and one forward pass returns the decision with calibrated probabilities - in 77 ms on an M1 Max. GGUF builds (BF16 / Q80 / Q4KM) for LM Studio and llama.cpp: chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-GGUF Everything below is measured on data the model never trained on, with the probabilities exactly as the released weights produce them (no post-processing). - Calibrated out of the box. An ECE of 0.017 on 1,500 examples is statistically indistinguishable from a perfectly calibrated model… Open weights apache-2.0 1.9B parameters 262,144 tokens mlx [View model](https://savrn.com/models/jev-style-qwen3-5-2b-decision-mlx-bf16) Model · Text generation ### [Jev-Style-Qwen3.5-2B-Decision-v2-MLX-bf16](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-mlx-bf16) [Chaoliang Yan](https://savrn.com/model-publishers/chaoliangunsw) Website: jevstyle.com — all JevStyle decision models, benchmarks and quickstart in one place. A Jev-style decision model for classification, routing and typed choices. Give it a state, a question and a list of options; one prefill returns a selected option with calibrated probabilities. Download this build: model.safetensors. The repository also includes its calibration, inference client and evaluation records. 81.20% macro accuracy on the fixed English reference panel, compared with 76.68% for v1 and 75.09% for English Laya. The results below use the CUDA reference structure: 11 real-label task groups, 3,277 decisions, equal task weights, and the same 3,100-record calibration split.… Open weights apache-2.0 1.9B parameters 262,144 tokens mlx [View model](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2-mlx-bf16) Model · Text generation ### [dQwen3.5-2B-Base](https://savrn.com/models/dqwen3-5-2b-base) [IFML](https://savrn.com/model-publishers/ut-ifml) A masked diffusion language model adapted from Qwen3.5-2B. The backbone is hybrid: only its attention layers are made bidirectional, and the Gated DeltaNet layers stay causal. This is a base model, with no instruction tuning. Paper: dQwen3.5: Hybrid-Attention Diffusion Language Models. Code: https://github.com/AntonXue/dQwen Needs a CUDA GPU and transformers>=5.13 (tested with torch 2.7.1+cu128, flash-linear-attention 0.5.1). generate decodes the whole canvas at once, committing positions above a confidence threshold (tau=0.9); pass blocklength=32 for left-to-right block decoding, or tau=None, stepsperblock=k for a fixed budget. The 50B-token checkpoint from the paper is… Open weights apache-2.0 1.9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/dqwen3-5-2b-base) Model · Text generation ### [vnpen-writer-2b-v0.1-preview](https://savrn.com/models/vnpen-writer-2b-v0-1-preview) [VNPen LLM Project](https://savrn.com/model-publishers/vnpen) VNPen is MewBaka Studio's visual-novel model series. The writer edition is for script writing, de-AI rewriting, and generating example scenes from a mood brief. Output format is one script line per line: speaker：text, with narration written as 旁白：. The base Qwen/Qwen3.5-2B is multimodal. Its checkpoint carries 297 model.visual. tensors (a depth-24 / hidden-1024 / patch-16 ViT) and 15 mtp. tensors for multi-token prediction. This project is text-only. In transformers, AutoModelForCausalLM on a qwen35 config builds Qwen35ForCausalLM over a Qwen35TextConfig — so the vision weights were never loaded at any point: not for training, not for merging, not for saving. Verified on the published… Open weights apache-2.0 1.9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/vnpen-writer-2b-v0-1-preview) Model · Text generation ### [ThinkLess-2B-FP8](https://savrn.com/models/thinkless-2b-fp8) [Sadikh Shaik](https://savrn.com/model-publishers/shaik1903) FP8 version of ThinkLess-2B: 8-bit floating-point weights and activations, 2.5 GB (bf16: 4.3 GB), with near-identical accuracy. Made with llm-compressor (FP8DYNAMIC: per-channel FP8 weights, dynamic per-token FP8 activations, no calibration data). The output head, vision tower and MTP heads stay in 16-bit. The differences are within the 95% confidence intervals, and answers stay just as short (cut-offs ≤ 1%). FP8 compute needs a GPU with FP8 support (NVIDIA Hopper or Ada, e.g. H100, L4, RTX 40-series); vLLM loads the compressed-tensors format directly. Use Qwen3.5's thinking-mode sampling (temperature 1.0, top-p 0.95, top-k 20, presence penalty 1.5). A 4-bit AWQ version of ThinkLess-2B was… Open weights apache-2.0 1.9B parameters 262,144 tokens transformers [View model](https://savrn.com/models/thinkless-2b-fp8) Model · Text generation ### [Shreyansh-STEM-AI-2B-v3](https://savrn.com/models/shreyansh-stem-ai-2b-v3) [Shreyansh singh](https://savrn.com/model-publishers/shreyansh12183) Shreyansh-STEM-AI-2B-v3 is a sovereign compact foundation model for scientific, physical, and mathematical derivation, engineered through SOLAR-style Depth Up-Scaling (DUS) and Continual Pre-Training (CPT) Seam Healing. To surpass standard 2B parameter capacity without requiring training from scratch, intermediate transformer layers were duplicated and spliced, expanding the model depth to 22 layers with a hidden dimension of $d=2048$. Splicing transformer blocks introduces interface discontinuity along the residual stream. To heal these seams, the model underwent Continual Pre-Training (CPT) over the multi-gigabyte shreyansh-1B-SLM-pretrain-stem-english corpus (over 2,400 parquet shards).… Open weights cc-by-nc-4.0 1.9B parameters 4,096 tokens [View model](https://savrn.com/models/shreyansh-stem-ai-2b-v3) ## Chaoliang Yan [All models and datasets](https://savrn.com/model-publishers/chaoliangunsw) ## Versions - [6b84dda83feb](https://savrn.com/models/jev-style-qwen3-5-2b-decision-v2/versions/6b84dda83feb) · current 2026-09-27 ## Explore More - [All text generation models](https://savrn.com/models/tasks/text-generation) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-09-27. - [Hugging Face record](https://huggingface.co/chaoliangUNSW/Jev-Style-Qwen3.5-2B-Decision-v2) - [How the hub is built](https://savrn.com/model-hub/methodology)