The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Add .bat file to nightly package to run with fp16 accumulation.
comfyanonymous4ab1875283ce
diff --git a/.ci/windows_nightly_base_files/run_nvidia_gpu_fast_fp16_accumulation.bat b/.ci/windows_nightly_base_files/run_nvidia_gpu_fast_fp16_accumulation.batnew file mode 100644index 000000000..38f06ecb2--- /dev/null+++ b/.ci/windows_nightly_base_files/run_nvidia_gpu_fast_fp16_accumulation.bat@@ -0,0 +1,2 @@+.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --fast fp16_accumulation+pause