vera-s1 0.1.2__tar.gz

vera_s1-0.1.2/PKG-INFO

@@ -0,0 +1,57 @@
+Metadata-Version: 2.4
+Name: vera-s1
+Version: 0.1.2
+Summary: vera-spark and vera-core: calibrated System One decision models
+License-Expression: Apache-2.0
+Project-URL: vera-spark, https://huggingface.co/scar-ai/vera-spark
+Project-URL: vera-core, https://huggingface.co/scar-ai/vera-core
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Requires-Dist: torch>=2.4
+Requires-Dist: transformers>=5.0
+Requires-Dist: huggingface_hub>=0.25
+Provides-Extra: fast
+Requires-Dist: flash-linear-attention; extra == "fast"
+
+# vera-s1
+
+Python package and Jev-compatible HTTP server for the **vera** System One decision models:
+[vera-spark](https://huggingface.co/scar-ai/vera-spark) (~630M, compact) and
+[vera-core](https://huggingface.co/scar-ai/vera-core) (~2B, most accurate).
+
+Give a vera model a state (text or JSON) and typed questions; it returns typed answers with calibrated probabilities
+in one forward pass. It never generates text.
+
+```bash
+pip install vera-s1            # add [fast] for flash-linear-attention (faster vera-core on GPU)
+```
+
+```python
+import vera
+
+agent = vera.load("scar-ai/vera-spark")       # or "scar-ai/vera-core"
+result = agent.predict(
+    "Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan.",
+    {"department": {"type": "choice", "instructions": "Which department should handle this?",
+                    "criteria": {"billing": "invoices, payments, refunds", "technical": "bugs, outages",
+                                 "other": "everything else"}},
+     "churn_risk": {"type": "noul", "instructions": "Does the user threaten to cancel or leave?"}})
+print(result["answers"])
+```
+
+Question types: `choice` (2 to 255 options), `score` (2 to 10 ordered levels) and `noul` (yes/no probability).
+
+## HTTP server
+
+```bash
+vera-serve --model scar-ai/vera-core --port 8200
+curl -s localhost:8200/v1/systemone -H 'Content-Type: application/json' -d '{
+  "state": "I was charged twice. Please fix this ASAP.",
+  "questions": {"billing": {"type": "noul", "instructions": "Is this ticket about billing?"}}}'
+```
+
+Same request and response shape as Jev's `POST /v1/systemone`. Binds `127.0.0.1` by default and has no
+authentication: put it behind your own gateway before exposing it.
+
+Requires Python 3.10+ and transformers 5+. A GPU (CUDA or ROCm) is recommended; Apple MPS is not supported (CPU is
+used on Macs).

vera_s1-0.1.2/README.md

@@ -0,0 +1,42 @@
+# vera-s1
+
+Python package and Jev-compatible HTTP server for the **vera** System One decision models:
+[vera-spark](https://huggingface.co/scar-ai/vera-spark) (~630M, compact) and
+[vera-core](https://huggingface.co/scar-ai/vera-core) (~2B, most accurate).
+
+Give a vera model a state (text or JSON) and typed questions; it returns typed answers with calibrated probabilities
+in one forward pass. It never generates text.
+
+```bash
+pip install vera-s1            # add [fast] for flash-linear-attention (faster vera-core on GPU)
+```
+
+```python
+import vera
+
+agent = vera.load("scar-ai/vera-spark")       # or "scar-ai/vera-core"
+result = agent.predict(
+    "Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan.",
+    {"department": {"type": "choice", "instructions": "Which department should handle this?",
+                    "criteria": {"billing": "invoices, payments, refunds", "technical": "bugs, outages",
+                                 "other": "everything else"}},
+     "churn_risk": {"type": "noul", "instructions": "Does the user threaten to cancel or leave?"}})
+print(result["answers"])
+```
+
+Question types: `choice` (2 to 255 options), `score` (2 to 10 ordered levels) and `noul` (yes/no probability).
+
+## HTTP server
+
+```bash
+vera-serve --model scar-ai/vera-core --port 8200
+curl -s localhost:8200/v1/systemone -H 'Content-Type: application/json' -d '{
+  "state": "I was charged twice. Please fix this ASAP.",
+  "questions": {"billing": {"type": "noul", "instructions": "Is this ticket about billing?"}}}'
+```
+
+Same request and response shape as Jev's `POST /v1/systemone`. Binds `127.0.0.1` by default and has no
+authentication: put it behind your own gateway before exposing it.
+
+Requires Python 3.10+ and transformers 5+. A GPU (CUDA or ROCm) is recommended; Apple MPS is not supported (CPU is
+used on Macs).

vera_s1-0.1.2/pyproject.toml

@@ -0,0 +1,25 @@
+[build-system]
+requires = ["setuptools>=77"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "vera-s1"
+version = "0.1.2"
+description = "vera-spark and vera-core: calibrated System One decision models"
+readme = "README.md"
+license = "Apache-2.0"
+requires-python = ">=3.10"
+dependencies = ["torch>=2.4", "transformers>=5.0", "huggingface_hub>=0.25"]
+
+[project.optional-dependencies]
+fast = ["flash-linear-attention"]
+
+[project.urls]
+"vera-spark" = "https://huggingface.co/scar-ai/vera-spark"
+"vera-core" = "https://huggingface.co/scar-ai/vera-core"
+
+[project.scripts]
+vera-serve = "vera.serve:main"
+
+[tool.setuptools]
+packages = ["vera"]

vera_s1-0.1.2/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

vera_s1-0.1.2/vera/__init__.py

@@ -0,0 +1,47 @@
+"""vera: calibrated System One decision models (vera-spark, vera-core).
+
+    import vera
+    agent = vera.load("scar-ai/vera-spark")
+    result = agent.predict(state, questions)
+"""
+from __future__ import annotations
+
+import os
+
+__version__ = "0.1.2"
+__all__ = ["load", "Agent"]
+
+
+class Agent:
+    """A loaded model. `predict` answers every question in one forward pass."""
+
+    def __init__(self, pred):
+        self.pred = pred
+
+    @property
+    def name(self) -> str:
+        return self.pred.name
+
+    @property
+    def device(self) -> str:
+        return self.pred.device
+
+    def predict(self, state, questions: dict) -> dict:
+        """state: text or any JSON-serialisable object. questions: {id: {"type": "choice"|"score"|"noul", ...}}.
+        Returns {"model", "answers": {id: {...}}, "usage"} (the same shape as POST /v1/systemone)."""
+        return self.pred.answer({"state": state, "questions": questions})
+
+    __call__ = predict
+
+
+def load(model: str = "scar-ai/vera-spark", device: str | None = None, max_len: int | None = None,
+         token: str | None = None, revision: str | None = None) -> Agent:
+    """Load a vera model from a Hub repo id (downloaded once and cached) or a local directory.
+    device: "cuda" or "cpu"; default is CUDA when available, else CPU."""
+    os.environ.setdefault("USE_TF", "0")  # transformers' TensorFlow probe can hang model construction
+    if not os.path.isdir(model):
+        from huggingface_hub import snapshot_download
+        model = snapshot_download(model, token=token, revision=revision,
+                                  allow_patterns=["model.pt", "arch.json", "tokenizer*", "chat_template.jinja"])
+    from .serve import Predictor
+    return Agent(Predictor(model, device=device, max_len=max_len))