Text Classification
Transformers
Safetensors
English
qwen3_5_moe
image-text-to-text
system-one
typed-decisions
decision-model
calibrated-probabilities
jev
noul
choice
score
mixture-of-experts
vllm
Eval Results (legacy)
Eval Results
Instructions to use autotrust/AutoTrust-Decider-Pro-35B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use autotrust/AutoTrust-Decider-Pro-35B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="autotrust/AutoTrust-Decider-Pro-35B")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("autotrust/AutoTrust-Decider-Pro-35B") model = AutoModelForMultimodalLM.from_pretrained("autotrust/AutoTrust-Decider-Pro-35B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Model card: NVFP4 version autotrust/JEV-35B-NVFP4
Browse files
README.md
CHANGED
|
@@ -55,6 +55,17 @@ model-index:
|
|
| 55 |
**72.08** (our rebuild of the vision benchmarks; JEV-27B-VL 71.67 on the same rebuild, a tie within noise). Median single-request latency
|
| 56 |
241 ms on one B200 (JEV-27B-VL 271 ms).
|
| 57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
> **Two models, two organisations.** **TypeSafe Jev 1.13** is the hosted, closed model made by TypeSafe AI.
|
| 59 |
> **autotrust/JEV-35B** is an independent open-weights model built by AutoTrust AI; it is not affiliated with,
|
| 60 |
> endorsed by, or a product of TypeSafe AI.
|
|
@@ -243,6 +254,14 @@ hf download autotrust/JEV-35B --local-dir JEV-35B
|
|
| 243 |
bash JEV-35B/serve.sh # vLLM on :8000; one GPU with 80 GB or more
|
| 244 |
```
|
| 245 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 246 |
`serve.sh` runs `serve_decide.py`: the standard vLLM OpenAI server (same flags as `vllm serve`) with a `POST /v1/decide`
|
| 247 |
route. Plain requests are System 2; requests for the LoRA module `jev-decision` (`adapter_vllm/`: backbone LoRA + the
|
| 248 |
decision head as an `lm_head` LoRA) are System 1. Tested with a vLLM development build from September 2026.
|
|
@@ -328,6 +347,8 @@ videos/ computer-use and robot-arm episodes
|
|
| 328 |
reports/demos/ per-episode computer-use, robot-arm and game results
|
| 329 |
```
|
| 330 |
|
|
|
|
|
|
|
| 331 |
## License
|
| 332 |
|
| 333 |
Apache-2.0. This repository contains the weights of [Qwen/Qwen3.5-35B-A3B](https://huggingface.co/Qwen/Qwen3.5-35B-A3B)
|
|
|
|
| 55 |
**72.08** (our rebuild of the vision benchmarks; JEV-27B-VL 71.67 on the same rebuild, a tie within noise). Median single-request latency
|
| 56 |
241 ms on one B200 (JEV-27B-VL 271 ms).
|
| 57 |
|
| 58 |
+
> [!TIP]
|
| 59 |
+
> **NVFP4 version: [autotrust/JEV-35B-NVFP4](https://huggingface.co/autotrust/JEV-35B-NVFP4)**, the same model with its
|
| 60 |
+
> routed experts in NVFP4: **25.6 GB download, 23 GiB of GPU memory** (bf16: 72 GB, 66 GiB), so it runs on one 32–48 GB
|
| 61 |
+
> GPU. Decision Index 0.3 public 59.18 (bf16 59.75), vision rebuild 71.23 (bf16 72.08), 96.6 % of the Decision Index
|
| 62 |
+
> answers identical to bf16, same latency, same `serve.sh` and API.
|
| 63 |
+
|
| 64 |
+
| version | download | GPU memory (weights) | Decision Index 0.3 public | vision (rebuild) |
|
| 65 |
+
|---|---:|---:|---:|---:|
|
| 66 |
+
| [autotrust/JEV-35B](https://huggingface.co/autotrust/JEV-35B) (bf16, this repo) | 72 GB | 66.5 GiB | 59.75 | 72.08 |
|
| 67 |
+
| [**autotrust/JEV-35B-NVFP4**](https://huggingface.co/autotrust/JEV-35B-NVFP4) | **25.6 GB** | **23.3 GiB** | 59.18 | 71.23 |
|
| 68 |
+
|
| 69 |
> **Two models, two organisations.** **TypeSafe Jev 1.13** is the hosted, closed model made by TypeSafe AI.
|
| 70 |
> **autotrust/JEV-35B** is an independent open-weights model built by AutoTrust AI; it is not affiliated with,
|
| 71 |
> endorsed by, or a product of TypeSafe AI.
|
|
|
|
| 254 |
bash JEV-35B/serve.sh # vLLM on :8000; one GPU with 80 GB or more
|
| 255 |
```
|
| 256 |
|
| 257 |
+
On a smaller GPU, use [autotrust/JEV-35B-NVFP4](https://huggingface.co/autotrust/JEV-35B-NVFP4) (one GPU with 32 GB or
|
| 258 |
+
more; same `serve.sh` and API):
|
| 259 |
+
|
| 260 |
+
```bash
|
| 261 |
+
hf download autotrust/JEV-35B-NVFP4 --local-dir JEV-35B-NVFP4
|
| 262 |
+
bash JEV-35B-NVFP4/serve.sh
|
| 263 |
+
```
|
| 264 |
+
|
| 265 |
`serve.sh` runs `serve_decide.py`: the standard vLLM OpenAI server (same flags as `vllm serve`) with a `POST /v1/decide`
|
| 266 |
route. Plain requests are System 2; requests for the LoRA module `jev-decision` (`adapter_vllm/`: backbone LoRA + the
|
| 267 |
decision head as an `lm_head` LoRA) are System 1. Tested with a vLLM development build from September 2026.
|
|
|
|
| 347 |
reports/demos/ per-episode computer-use, robot-arm and game results
|
| 348 |
```
|
| 349 |
|
| 350 |
+
NVFP4 weights (routed experts in NVFP4, everything else identical): [autotrust/JEV-35B-NVFP4](https://huggingface.co/autotrust/JEV-35B-NVFP4).
|
| 351 |
+
|
| 352 |
## License
|
| 353 |
|
| 354 |
Apache-2.0. This repository contains the weights of [Qwen/Qwen3.5-35B-A3B](https://huggingface.co/Qwen/Qwen3.5-35B-A3B)
|