AshScholar/r1m-Q4_K_M-GGUF

This model was converted to GGUF format from AshScholar/r1m using llama.cpp via the ggml.ai's GGUF-my-repo space.

About Model

This model was finetuned from Qwen2.5-1M on data from DeepSeek R1. This model uses CoT, and also decides when or when not to use CoT. Credits to Rombo Org for the curated dataset from R1. r1m has a one million token context and near R1 performance.

Specs

7 billion parameters 1 million token context Utilizes tokens.

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo AshScholar/r1m-Q4_K_M-GGUF --hf-file r1m-q4_k_m.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo AshScholar/r1m-Q4_K_M-GGUF --hf-file r1m-q4_k_m.gguf -c 2048

Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.

Step 1: Clone llama.cpp from GitHub.

git clone https://github.com/ggerganov/llama.cpp

Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).

cd llama.cpp && LLAMA_CURL=1 make

Step 3: Run inference through the main binary.

./llama-cli --hf-repo AshScholar/r1m-Q4_K_M-GGUF --hf-file r1m-q4_k_m.gguf -p "The meaning to life and the universe is"

./llama-server --hf-repo AshScholar/r1m-Q4_K_M-GGUF --hf-file r1m-q4_k_m.gguf -c 2048