programasweights 0.3.0__tar.gz → 0.3.2__tar.gz

{programasweights-0.3.0 → programasweights-0.3.2}/AGENTS.md

@@ -20,7 +20,7 @@ import programasweights as paw
 # "email-triage" is an official pre-compiled program (slug)
 fn = paw.function("email-triage")
 fn("Urgent: server is down!")  # "immediate"
-fn("Newsletter: spring picnic")  # "can wait"
+fn("Newsletter: spring picnic")  # "wait"
 
 # Compile your own from a description
 program = paw.compile(
@@ -28,7 +28,7 @@ program = paw.compile(
     compiler="paw-4b-qwen3-0.6b"  # or "paw-4b-gpt2" for smaller/faster
 )
 fn = paw.function(program.id)
-fn("{name: 'Alice',}")  # '{"name": "Alice"}'
+fn("{name: 'Alice',}")  # '{"name":"Alice"}'
 
 # Or compile and load in one step
 fn = paw.compile_and_load("Classify sentiment as positive or negative")
@@ -38,7 +38,7 @@ fn("I love this!")  # "positive"
 ## Two Compilers
 
 - **Standard** (`paw-4b-qwen3-0.6b`) — higher accuracy, 594 MB base + ~22 MB/program. Default.
-- **Compact** (`paw-4b-gpt2`) — smaller (105 MB base + ~5 MB/program), runs in browser via WebAssembly.
+- **Compact** (`paw-4b-gpt2`) — smaller (134 MB base + ~5 MB/program), runs in browser via WebAssembly.
 
 ## When to Use PAW
 
@@ -211,10 +211,12 @@ paw.login()
 | `RuntimeError: assets not ready` on download | Program still generating after compile | SDK polls automatically for up to 30s. If persistent, recompile. |
 | `httpx.HTTPStatusError: 422` on compile | Spec too short (<10 chars) | Adjust spec length. |
 | `httpx.HTTPStatusError: 429` | Rate limit exceeded | Wait, or sign in for higher limits. |
+| GPU/Metal errors on load | GPU backend not available or incompatible | Set `PAW_GPU_LAYERS=0` or pass `n_gpu_layers=0` to force CPU. |
 
 ## Performance
 
-- **First call** ~5-15s (loads base model). Subsequent calls 0.5-5s depending on input length.
+- **GPU acceleration** enabled by default (`n_gpu_layers=-1`). Uses Metal on Mac, CUDA on Linux, falls back to CPU automatically. If GPU causes issues, set `PAW_GPU_LAYERS=0` or pass `n_gpu_layers=0`.
+- **First call** ~1-5s (loads base model). Subsequent calls ~0.05-0.5s depending on input length and GPU availability.
 - **Base model shared** across functions on disk. Each LoRA adapter adds ~22 MB.
 - **Cache**: `~/.cache/programasweights/`. Override with `PAW_CACHE_DIR`.
 - **Offline** after first download.

{programasweights-0.3.0 → programasweights-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: programasweights
-Version: 0.3.0
+Version: 0.3.2
 Summary: Compile natural language specifications into neural programs that run locally via llama.cpp.
 Project-URL: Homepage, https://programasweights.com
 Project-URL: Repository, https://github.com/programasweights/programasweights-python
@@ -48,7 +48,7 @@ import programasweights as paw
 # Use a pre-compiled function (downloads once, runs locally forever)
 fn = paw.function("email-triage")
 fn("Urgent: the server is down!")        # "immediate"
-fn("Newsletter: spring picnic")          # "can wait"
+fn("Newsletter: spring picnic")          # "wait"
 
 # Compile your own from a description
 program = paw.compile(
@@ -57,7 +57,7 @@ program = paw.compile(
     slug="json-fixer"              # optional: creates username/json-fixer handle
 )
 fn = paw.function(program.slug)    # or paw.function(program.id)
-fn("{name: 'Alice',}")  # '{"name": "Alice"}'
+fn("{name: 'Alice',}")  # '{"name":"Alice"}'
 
 # Or compile and load in one step
 fn = paw.compile_and_load("Classify sentiment as positive or negative")
@@ -66,17 +66,20 @@ fn("I love this!")  # "positive"
 
 ## Two Compilers
 
-|                    | Standard (Qwen3 0.6B)   | Compact (GPT-2 124M)   |
-|--------------------|-------------------------|------------------------|
-| Compiler name      | `paw-4b-qwen3-0.6b`    | `paw-4b-gpt2`          |
-| Accuracy           | Higher                  | Lower                  |
-| Base model size    | 594 MB                  | 105 MB                 |
-| Program size       | ~22 MB                  | ~5 MB                  |
-| Inference speed    | ~90ms (server)          | ~50ms (server)         |
-| Runs in browser    | No                      | Yes                    |
+
+|                 | Standard (Qwen3 0.6B) | Compact (GPT-2 124M) |
+| --------------- | --------------------- | -------------------- |
+| Compiler name   | `paw-4b-qwen3-0.6b`   | `paw-4b-gpt2`        |
+| Accuracy        | Higher                | Lower                |
+| Base model size | 594 MB                | 134 MB               |
+| Program size    | ~22 MB                | ~5 MB                |
+| Local inference | ~0.05-0.5s per call   | ~0.03-0.3s per call  |
+| Runs in browser | No                    | Yes (WebAssembly)    |
 
 Default is Standard (Qwen3 0.6B). Use Compact (GPT-2) when you need smaller files or browser deployment.
 
+GPU acceleration is enabled by default (Metal on Mac, CUDA on Linux, falls back to CPU). Set `PAW_GPU_LAYERS=0` to force CPU if GPU causes issues.
+
 ## Browser SDK
 
 Programs compiled with GPT-2 also run entirely in the browser via WebAssembly — no server needed, data never leaves the user's device.
@@ -99,9 +102,9 @@ See the [browser SDK repo](https://github.com/programasweights/programasweights-
 
 PAW works with Cursor, Claude, Codex, and other AI coding assistants. Paste this into your agent's chat:
 
-> I want to use ProgramAsWeights (PAW) to create fuzzy text functions that run locally. Read the instructions at https://programasweights.com/agents and help me integrate it.
+> I want to use ProgramAsWeights (PAW) to create fuzzy text functions that run locally. Read the instructions at [https://programasweights.com/AGENTS.md](https://programasweights.com/AGENTS.md) and help me integrate it.
 
-Or save [`AGENTS.md`](https://programasweights.com/agents) to your project root — agents read it automatically.
+Or save `[AGENTS.md](https://programasweights.com/agents)` to your project root — agents read it automatically.
 
 ## When to Use PAW
 
@@ -145,4 +148,4 @@ paw login
 
 ## License
 
-MIT
+MIT

{programasweights-0.3.0 → programasweights-0.3.2}/PYPI_README.md

@@ -18,7 +18,7 @@ import programasweights as paw
 # Use a pre-compiled function (downloads once, runs locally forever)
 fn = paw.function("email-triage")
 fn("Urgent: the server is down!")        # "immediate"
-fn("Newsletter: spring picnic")          # "can wait"
+fn("Newsletter: spring picnic")          # "wait"
 
 # Compile your own from a description
 program = paw.compile(
@@ -27,7 +27,7 @@ program = paw.compile(
     slug="json-fixer"              # optional: creates username/json-fixer handle
 )
 fn = paw.function(program.slug)    # or paw.function(program.id)
-fn("{name: 'Alice',}")  # '{"name": "Alice"}'
+fn("{name: 'Alice',}")  # '{"name":"Alice"}'
 
 # Or compile and load in one step
 fn = paw.compile_and_load("Classify sentiment as positive or negative")
@@ -36,17 +36,20 @@ fn("I love this!")  # "positive"
 
 ## Two Compilers
 
-|                    | Standard (Qwen3 0.6B)   | Compact (GPT-2 124M)   |
-|--------------------|-------------------------|------------------------|
-| Compiler name      | `paw-4b-qwen3-0.6b`    | `paw-4b-gpt2`          |
-| Accuracy           | Higher                  | Lower                  |
-| Base model size    | 594 MB                  | 105 MB                 |
-| Program size       | ~22 MB                  | ~5 MB                  |
-| Inference speed    | ~90ms (server)          | ~50ms (server)         |
-| Runs in browser    | No                      | Yes                    |
+
+|                 | Standard (Qwen3 0.6B) | Compact (GPT-2 124M) |
+| --------------- | --------------------- | -------------------- |
+| Compiler name   | `paw-4b-qwen3-0.6b`   | `paw-4b-gpt2`        |
+| Accuracy        | Higher                | Lower                |
+| Base model size | 594 MB                | 134 MB               |
+| Program size    | ~22 MB                | ~5 MB                |
+| Local inference | ~0.05-0.5s per call   | ~0.03-0.3s per call  |
+| Runs in browser | No                    | Yes (WebAssembly)    |
 
 Default is Standard (Qwen3 0.6B). Use Compact (GPT-2) when you need smaller files or browser deployment.
 
+GPU acceleration is enabled by default (Metal on Mac, CUDA on Linux, falls back to CPU). Set `PAW_GPU_LAYERS=0` to force CPU if GPU causes issues.
+
 ## Browser SDK
 
 Programs compiled with GPT-2 also run entirely in the browser via WebAssembly — no server needed, data never leaves the user's device.
@@ -69,9 +72,9 @@ See the [browser SDK repo](https://github.com/programasweights/programasweights-
 
 PAW works with Cursor, Claude, Codex, and other AI coding assistants. Paste this into your agent's chat:
 
-> I want to use ProgramAsWeights (PAW) to create fuzzy text functions that run locally. Read the instructions at https://programasweights.com/agents and help me integrate it.
+> I want to use ProgramAsWeights (PAW) to create fuzzy text functions that run locally. Read the instructions at [https://programasweights.com/AGENTS.md](https://programasweights.com/AGENTS.md) and help me integrate it.
 
-Or save [`AGENTS.md`](https://programasweights.com/agents) to your project root — agents read it automatically.
+Or save `[AGENTS.md](https://programasweights.com/agents)` to your project root — agents read it automatically.
 
 ## When to Use PAW
 
@@ -115,4 +118,4 @@ paw login
 
 ## License
 
-MIT
+MIT

{programasweights-0.3.0 → programasweights-0.3.2}/docs/adr/001-llama-cpp-over-pytorch.md

@@ -17,5 +17,5 @@ Replace the PyTorch runtime with llama-cpp-python (~80MB install). Use GGUF mode
 - Inference uses Metal (Mac), CPU (Linux/Windows) — no GPU required
 - Must pre-render chat templates server-side (no transformers tokenizer on client)
 - .paw format must change from v2 (safetensors) to v3 (GGUF adapter)
-- Base model is downloaded once (~594 MB for Q6_K) and shared across all functions
+- Base model is downloaded once (~594 MB Q6_K for Qwen3, ~134 MB Q8_0 for GPT-2) and shared across all functions
 - Per-function adapter download is ~23MB (Q4_0, confirmed lossless at 4096-scale eval)

{programasweights-0.3.0 → programasweights-0.3.2}/docs/advanced/adrs.md

@@ -14,16 +14,17 @@ Concise records of major technical choices. Full ADR files may live elsewhere in
 
 ---
 
-## ADR 002: Q4_0 adapter and Q6_K base quantization
+## ADR 002: Quantization levels for base models and adapters
 
-**Decision:** Use **Q4_0** for adapters and **Q6_K** for the base model in shipped bundles.
+**Decision:** Use **Q4_0** for adapters. For base models, use **Q6_K** for Qwen3 0.6B and **Q8_0** for GPT-2.
 
 **Context:** Empirical evaluation on **4096** held-out examples across quantization settings informed the trade-off.
 
 **Consequence:**
 
-- **Q6_K base** — quality is preserved while the footprint is roughly **60% smaller** than fp16.
-- **Q4_0 adapter** — quality loss is negligible; adapter size drops to about **23 MB** versus **78 MB** for a heavier format at comparable settings in prior experiments.
+- **Q6_K base (Qwen3 0.6B)** — quality is preserved while the footprint is roughly **60% smaller** than fp16. 4096-sample eval shows no accuracy loss vs fp16.
+- **Q8_0 base (GPT-2)** — Q6_K caused ~3.5% accuracy loss on GPT-2 (4096-sample eval). Q8_0 closes the gap at only ~29 MB additional cost (134 MB vs 105 MB).
+- **Q4_0 adapter** — quality loss is negligible for both models; adapter size drops to about **23 MB** (Qwen3) / **5 MB** (GPT-2) versus **78 MB** / **19 MB** at fp16.
 
 ---
 

{programasweights-0.3.0 → programasweights-0.3.2}/docs/guide/browser-inference.md

@@ -84,7 +84,7 @@ paw.configure({
 
 | Component | Size | When |
 |-----------|------|------|
-| Base model (GPT-2 Q6_K) | 105 MB | First program load (cached) |
+| Base model (GPT-2 Q8_0) | 134 MB | First program load (cached) |
 | LoRA adapter | ~5 MB | Per program |
 | Prefix cache | ~7 MB | Per program |
 | **First load total** | **~117 MB** | |

{programasweights-0.3.0 → programasweights-0.3.2}/docs/index.md

@@ -20,7 +20,7 @@ result = fn("Thesis defense committee needs your signature by EOD")
 print(result)  # "immediate"
 
 result = fn("Department newsletter: spring picnic next Friday")
-print(result)  # "can wait"
+print(result)  # "wait"
 ```
 
 ## Compile Your Own
@@ -35,7 +35,7 @@ program = paw.compile(
 
 # Use the compiled program
 fn = paw.function(program.id)
-fn("{name: 'Alice', age: 30,}")  # '{"name": "Alice", "age": 30}'
+fn("{name: 'Alice', age: 30,}")  # '{"name":"Alice","age":30}'
 ```
 
 ## What Can You Build?
@@ -56,7 +56,7 @@ Each compiled program has two parts:
 1. **Discrete pseudo-program** -- text instructions generated by the neural compiler
 2. **Continuous neural adapter** -- LoRA weights (~23 MB) that tune the interpreter model
 
-At runtime, the SDK loads a Q6_K base model (~594 MB, downloaded once) and applies the LoRA adapter. Inference runs entirely locally via llama.cpp.
+At runtime, the SDK loads a quantized base model (Q6_K ~594 MB for Qwen3, Q8_0 ~134 MB for GPT-2, downloaded once) and applies the LoRA adapter. Inference runs entirely locally via llama.cpp.
 
 ## Browser Inference
 

{programasweights-0.3.0 → programasweights-0.3.2}/programasweights/__init__.py

@@ -12,7 +12,7 @@ Quick start:
     # Compile your own from a description
     program = paw.compile("Fix malformed JSON: repair missing quotes and trailing commas")
     fn = paw.function(program.id)
-    fn("{name: 'Alice',}")  # '{"name": "Alice"}'
+    fn("{name: 'Alice',}")  # '{"name":"Alice"}'
 
 API reference:
     paw.compile(spec)          Compile a spec on the server, returns Program
@@ -27,7 +27,7 @@ try:
     from importlib.metadata import version as _meta_version
     __version__ = _meta_version("programasweights")
 except Exception:
-    __version__ = "0.3.0"
+    __version__ = "0.3.1"
 
 from .config import get_api_url, get_api_key, set_api_key
 
@@ -39,6 +39,7 @@ def compile(
     tags: list[str] | None = None,
     public: bool = True,
     slug: str | None = None,
+    ephemeral: bool = False,
 ):
     """Compile a natural language specification into a neural program.
 
@@ -65,13 +66,13 @@ def compile(
         ... )
         >>> fn = paw.function(program.slug)  # or paw.function(program.id)
         >>> fn("{name: 'Alice',}")
-        '{"name": "Alice"}'
+        '{"name":"Alice"}'
     """
     from .client import PAWClient
     from ._output import status
     status("Compiling...")
     client = PAWClient(api_url=get_api_url(), api_key=get_api_key())
-    result = client.compile(spec, compiler=compiler, name=name, tags=tags, public=public, slug=slug)
+    result = client.compile(spec, compiler=compiler, name=name, tags=tags, public=public, slug=slug, ephemeral=ephemeral)
     label = f"{result.id}"
     if result.slug:
         ver_str = f" v{result.version}" if result.version and result.version > 1 else ""
@@ -101,8 +102,9 @@ def function(
         program_id: Program ID (str), slug (``da03/my-program``), pinned version
             (``da03/my-program@v3``), or a ``Program`` object from compile().
         n_ctx: Context window size for llama.cpp.
-        n_gpu_layers: GPU layers (-1 = all, 0 = CPU only). Defaults to CPU.
-            Set ``PAW_GPU_LAYERS`` env var or pass explicitly for GPU acceleration.
+        n_gpu_layers: GPU layers (-1 = all GPU, 0 = CPU only). Defaults to -1
+            (auto-uses Metal/CUDA if available, safe fallback to CPU).
+            Set ``PAW_GPU_LAYERS=0`` env var to force CPU-only.
         verbose: Print llama.cpp debug output.
         offline: Skip server check for slug resolution and use local cache only.
             Also set via ``PAW_OFFLINE=1`` env var.
@@ -125,7 +127,7 @@ def function(
     from .runtime_llamacpp import PawFunction
 
     if n_gpu_layers is None:
-        n_gpu_layers = int(os.environ.get("PAW_GPU_LAYERS", "0"))
+        n_gpu_layers = int(os.environ.get("PAW_GPU_LAYERS", "-1"))
 
     from ._output import status
 

{programasweights-0.3.0 → programasweights-0.3.2}/programasweights/cache.py

@@ -5,6 +5,7 @@ Cache structure:
     ~/.cache/programasweights/
         base_models/
             qwen3-0.6b-q6_k.gguf      # ~594 MB, downloaded once
+            gpt2-q8_0.gguf            # ~134 MB, downloaded once
         programs/
             <program_id>/
                 adapter.gguf            # ~23 MB, Q4_0 LoRA
@@ -25,12 +26,12 @@ from . import config
 
 BASE_MODEL_URLS = {
     "qwen3-0.6b-q6_k": "https://huggingface.co/programasweights/Qwen3-0.6B-GGUF-Q6_K/resolve/main/qwen3-0.6b-q6_k.gguf",
-    "gpt2-q6_k": "https://huggingface.co/programasweights/GPT2-GGUF-Q6_K/resolve/main/gpt2-q6_k.gguf",
+    "gpt2-q8_0": "https://huggingface.co/programasweights/GPT2-GGUF-Q8_0/resolve/main/gpt2-q8_0.gguf",
 }
 
 INTERPRETER_TO_GGUF = {
     "Qwen/Qwen3-0.6B": "qwen3-0.6b-q6_k",
-    "gpt2": "gpt2-q6_k",
+    "gpt2": "gpt2-q8_0",
 }
 
 

{programasweights-0.3.0 → programasweights-0.3.2}/programasweights/client.py

@@ -52,6 +52,7 @@ class PAWClient:
         tags: list[str] | None = None,
         public: bool = True,
         slug: str | None = None,
+        ephemeral: bool = False,
     ) -> Program:
         """Compile a spec into a neural program on the server.
 
@@ -71,6 +72,8 @@ class PAWClient:
             httpx.HTTPStatusError: On API errors (422 for validation, 429 for rate limit).
         """
         body: dict = {"spec": spec, "compiler": compiler, "public": public}
+        if ephemeral:
+            body["ephemeral"] = True
         if name:
             body["name"] = name
         if tags:
@@ -121,7 +124,7 @@ class PAWClient:
         from ._output import status
 
         status(f"Downloading program {program_id[:12]}...")
-        max_wait = 30
+        max_wait = 60
         elapsed = 0
         resp = None
         waiting_logged = False

{programasweights-0.3.0 → programasweights-0.3.2}/programasweights/runtime_llamacpp.py

@@ -88,7 +88,6 @@ class PawFunction:
                 _os.close(_old_stderr)
 
         placeholder = "{INPUT_PLACEHOLDER}"
-        self._use_special = interpreter not in ("gpt2",)
 
         if placeholder in self._template:
             prefix_text = self._template.split(placeholder)[0]
@@ -99,8 +98,8 @@ class PawFunction:
 
         self._prefix_tokens = self._llm.tokenize(
             prefix_text.encode("utf-8"),
-            add_bos=not self._use_special,
-            special=self._use_special,
+            add_bos=False,
+            special=True,
         )
         self._suffix_text = suffix_text
         self._n_prefix = len(self._prefix_tokens)
@@ -183,7 +182,7 @@ class PawFunction:
         input_tokens = self._llm.tokenize(
             input_with_suffix.encode("utf-8"),
             add_bos=False,
-            special=self._use_special,
+            special=True,
         )
 
         tokens_used = self._n_prefix + len(input_tokens)

{programasweights-0.3.0 → programasweights-0.3.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "programasweights"
-version = "0.3.0"
+version = "0.3.2"
 description = "Compile natural language specifications into neural programs that run locally via llama.cpp."
 readme = "PYPI_README.md"
 requires-python = ">=3.9"