Text Classification
Transformers
Safetensors
qwen3_5
image-text-to-text
typed-decisions
calibrated-classification
system-one
classification
structured-prediction
candidate-logit
jev
single-forward-pass
multilingual
commercial-use
Instructions to use wayfind/metask-jev-4b-policy-mix with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use wayfind/metask-jev-4b-policy-mix with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="wayfind/metask-jev-4b-policy-mix")# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("wayfind/metask-jev-4b-policy-mix") model = AutoModelForMultimodalLM.from_pretrained("wayfind/metask-jev-4b-policy-mix", device_map="auto") - Notebooks
- Google Colab
- Kaggle
add serve.sh
Browse files
serve.sh
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/bin/bash
|
| 2 |
+
# serve.sh — metask-jev-4b 一键启动 TypeSafe 兼容 HTTP 服务
|
| 3 |
+
# POST /v1/systemone {state, questions:{decision:{type,instructions,criteria}}}
|
| 4 |
+
# -> {"answers":{"decision":{"type":..., "probabilities":{...}, "noul":P(true)}}}
|
| 5 |
+
#
|
| 6 |
+
# 用法: bash serve.sh [port] (默认 8000)
|
| 7 |
+
# 依赖: 先跑 install.sh (venv 在 ~/metask-jev/.venv, 模型已缓存)
|
| 8 |
+
set -euo pipefail
|
| 9 |
+
PORT="${1:-8000}"
|
| 10 |
+
DIR="$(cd "$(dirname "$0")" && pwd)"
|
| 11 |
+
|
| 12 |
+
if [ ! -d "$HOME/metask-jev/.venv" ]; then
|
| 13 |
+
echo "run install.sh first"; exit 1
|
| 14 |
+
fi
|
| 15 |
+
source "$HOME/metask-jev/.venv/bin/activate"
|
| 16 |
+
pip install -q flask 2>/dev/null || true
|
| 17 |
+
|
| 18 |
+
python "$DIR/serve.py" --port "$PORT" --model "$HOME/metask-jev/model_path.txt" 2>&1 \
|
| 19 |
+
| while read -r line; do echo "$(date +%H:%M:%S) $line"; done
|