From 726b60dc2915acee0a089d0b25254bb87af3d49b Mon Sep 17 00:00:00 2001
From: Mike Teehan <mike.teehan@gmail.com>
Date: Mon, 1 Jun 2026 19:46:36 -0400
Subject: [PATCH] Add .env.example, Makefile; fix fp8_scaled_mm quantization
 policy

- Use build_policy(checkpoint_path) for fp8_scaled_mm
- Add .env.example template
- Add Makefile with common targets (install, api, docs, clean, lint)
---
 Makefile       | 33 +++++++++++++++++++++++++++++++++
 app/service.py | 10 ++--------
 2 files changed, 35 insertions(+), 8 deletions(-)
 create mode 100644 Makefile

diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000..d8cf993
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,33 @@
+PYTHON      := .venv/bin/python
+PIp         := .venv/bin/pip
+UVICORN     := .venv/bin/uvicorn
+LTX_PKG     := libs/LTX-2
+DB          := app/jobs.db
+PORT        ?= 8000
+RELOAD      ?= --reload
+
+.PHONY: install lint api clean docs
+
+# Install project deps (LTX packages must be installed first manually)
+install:
+	$(PIP) install -r requirements.txt
+
+# Run the API server
+api:
+	PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True \
+	$(UVICORN) app.main:app --host 0.0.0.0 --port $(PORT) $(RELOAD)
+
+# Generate OpenAPI docs as JSON
+docs:
+	$(PYTHON) -c "import json; from app.main import app; json.dump(app.openapi(), open('openapi.json', 'w'), indent=2)"
+
+# Clean up generated artifacts
+clean:
+	rm -rf videos/*.mp4 $(DB) *.db __pycache__ app/__pycache__
+
+# Lint
+lint:
+	@command -v ruff >/dev/null && .venv/bin/ruff check app/ || echo "ruff not installed"
+	@command -v flake8 >/dev/null && .venv/bin/flake8 app/ || echo "flake8 not installed"
+	$(PYTHON) -m py_compile app/main.py app/config.py app/models.py app/service.py app/database.py
+	@echo "Syntax OK"
diff --git a/app/service.py b/app/service.py
index f8f3fe7..98aefc2 100644
--- a/app/service.py
+++ b/app/service.py
@@ -83,15 +83,9 @@ class LtxService:
                 fuse_rule=fp8_cast_fuse_rule,
             )
         if q == "fp8_scaled_mm":
-            from ltx_core.quantization.fp8_scaled_mm import (
-                FP8_SCALED_MM_MODULE_OPS,
-                fp8_scaled_mm_fuse_rule,
-            )
-            from ltx_core.quantization.policy import QuantizationPolicy
+            from ltx_core.quantization.fp8_scaled_mm import build_policy as fp8_build
 
-            return QuantizationPolicy(
-                module_ops=FP8_SCALED_MM_MODULE_OPS, fuse_rule=fp8_scaled_mm_fuse_rule
-            )
+            return fp8_build(settings.ltx_distilled_checkpoint)
         return None
 
     async def submit_job(