ddpasa
ddpasa
As of 10 days ago: https://github.com/ggerganov/llama.cpp/commit/2307523d322af762ae06648b29ec5a9eb1c73032 This is great news for people who non-CUDA cards. What's necessary to support this with Ollama? I'm happy to help if you show me...
I'm coming here from ollama, where I'm trying to expose llama.cpp's vulkan backend in ollama (https://github.com/ollama/ollama/pull/2578). It works fine for models like phi-2 and tinyllama, but multimodal models such as...
I think Intel CPUs/GPUs now support more efficient inference with OpenVINO. See example here with LLAVA: https://docs.openvino.ai/2023.2/notebooks/257-llava-multimodal-chatbot-with-output.html It would be great if ollama could automatically default to OpenVINO on Intel...
This is a very preliminary ~~implementation~~ hack of Vulkan support, which llama.cpp recently added. This is not intended to be merged. This code is far from there. I just want...
Yi LM [is supported in ollama](https://ollama.ai/library/yi), but I don't think the multimodel Yi-VL models are. These are supposed to be very good, so it would be great to have them....
I have a intel CPU that supports a number of AVX features, but most of them are not picked up when using ollama. Below is the llama.log file: system info:...
In Turkey it's quite common for houses and buildings to have a name as well as a number. See https://taginfo.openstreetmap.org/keys/addr%3Ahousename for the corresponding taginfo.
### What model would you like? MobileVLM v2 is a very promising multimodal model that is already supported by llama.cpp. Here are the 3 versions: 1.7b: https://huggingface.co/mtgv/MobileVLM_V2-1.7B 3b: https://huggingface.co/mtgv/MobileVLM_V2-3B 7b:...
OpenELM are a series of small language models from Apple: https://huggingface.co/apple/OpenELM
In the past the UI allowed us to set any file extention for the caption files. Now it's a dropdown box with just 3 options. Please restore the older flexibility,...