[CI] use Q2_K Command-R model to reduce disk usage on CI

Signed-off-by: dm4 <dm4@secondstate.io>

Author: dm4

.github/workflows/llama.yml

@@ -119,11 +119,11 @@ jobs:
             run: |
               test -f ~/.wasmedge/env && source ~/.wasmedge/env
               cd wasmedge-ggml/command-r
-              curl -LO https://huggingface.co/andrewcanis/c4ai-command-r-v01-GGUF/resolve/main/c4ai-command-r-v01-Q5_K_M.gguf
+              curl -LO https://huggingface.co/andrewcanis/c4ai-command-r-v01-GGUF/resolve/main/c4ai-command-r-v01-Q2_K.gguf
               cargo build --target wasm32-wasi --release
               time wasmedge --dir .:. \
                 --env n_gpu_layers="$NGL" \
-                --nn-preload default:GGML:AUTO:c4ai-command-r-v01-Q5_K_M.gguf \
+                --nn-preload default:GGML:AUTO:c4ai-command-r-v01-Q2_K.gguf \
                 target/wasm32-wasi/release/wasmedge-ggml-command-r.wasm \
                 default \
                 '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>What is the capital of the United States?<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>'