media Hugging Face Forums · 1h ago · open_models

Celestium Neon fixes SDXL OOM on 4GB GPUs by cleaning VRAM

from English

Celestium Neon is a new tool designed to clean VRAM and enable Stable Diffusion XL (SDXL) inference on graphics cards with only 4GB of memory, such as the GTX 1650.

The tool operates through several key mechanisms:

L1-L2 PyTorch Cleanup to free VRAM from zombie tensors
Real-time VRAM monitoring to display MiB drops live
SDXL RESET functionality to force cleanup when out-of-memory errors occur
Automatic cleanup intervals set every 5 minutes

Testing on a GTX 1650 with driver 591.86 demonstrated that the tool successfully prevents OOM crashes, reducing VRAM usage to 938 MiB after cleanup and allowing image generation to complete.

Importance 1/3 Trust 1/3 Hugging Face Forums Inference efficiency Local inference

Read original

github llama.cpp · 4h ago · 1 view

llama.cpp b10636 disables npm UI build by default and cleans up CI

The llama.cpp project released version b10636, which changes the default behavior for building the web interface. The npm-based UI build is now disabled by default (LLAMA_BUILD_UI=OFF), with the UI instead embedded from local sources or prebuilt downloads.

github llama.cpp · 5h ago · 1 view

llama.cpp b10635 unblocks mmq for MoE on sm_60

The llama.cpp project released version b10635, which includes specific updates to CUDA support. The primary change is the unblocking of mixed-precision matrix multiplication (mmq) for Mixture of Experts (MoE) architectures on compute capability 6.0 (sm_60) GPUs.

github llama.cpp · 2d ago · 11 views

llama.cpp b10636 disables npm UI build by default and cleans up CI

llama.cpp b10635 unblocks mmq for MoE on sm_60

llama.cpp b10604 adds Deepseek 4 DSV4 support and macOS/iOS binaries

llama.cpp b10598 uses pillow-accurate resize algo and corrects resize_algo for all models

llama.cpp b10593 fixes DeepseekV4 rollback and model loading