llava image describe

bin/llava-describe.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+
+SIZE=7b
+
+LLAVA_PROMPT=$(
+cat <<EOF
+Generate a JSON description of the image by adapting the following JSON template:
+
+{
+    extended_description: ""
+    number_of_people: ,   
+    objects_in_picture_list: [
+        ""
+    ]
+}
+EOF
+)
+
+if [ "$SIZE" == "7b" ]; then
+    GGML_MODEL=~/.ai/models/llava/mys/ggml_llava-v1.5-7b/ggml-model-f16.gguf
+    MMPROJ_MODEL=~/.ai/models/llava/mys/ggml_llava-v1.5-7b/mmproj-model-f16.gguf
+else
+    GGML_MODEL=~/.ai/models/llava/mys/ggml_llava-v1.5-13b/ggml-model-q4_k.gguf
+    MMPROJ_MODEL=~/.ai/models/llava/mys/ggml_llava-v1.5-13b/mmproj-model-f16.gguf
+fi
+
+llava \
+    --prompt-cache "$HOME/.ai/cache/llava.cache" \
+    --model "$GGML_MODEL" \
+    --mmproj "$MMPROJ_MODEL" \
+    --image "$1" \
+    --prompt "$LLAVA_PROMPT" \
+    --n-gpu-layers 1 \
+    --log-disable 2>/dev/null | grep -v clip_model_load

install.sh

@@ -10,5 +10,6 @@ install -C -v ./bin/llama-finetune.sh ~/.local/bin
 install -C -v ./bin/llama.sh ~/.local/bin
 install -C -v ./bin/llama-menu.sh ~/.local/bin
 install -C -v ./bin/llama-server.sh ~/.local/bin
+install -C -v ./bin/llava-describe.sh ~/.local/bin
 
 echo "done"