diff --git a/Pipes.py b/Pipes.py index d4efd9b..f4c4ca0 100644 --- a/Pipes.py +++ b/Pipes.py @@ -1,7 +1,7 @@ import os import logging from dotenv import load_dotenv -from ezlocalai.LLM import LLM +from ezlocalai.LLM import LLM, is_vision_model from ezlocalai.STT import STT from ezlocalai.CTTS import CTTS from pyngrok import ngrok @@ -58,8 +58,10 @@ def __init__(self): else: logging.info(f"[LLM] {self.current_llm} model loading. Please wait...") self.llm = LLM(model=self.current_llm) + if is_vision_model(self.current_llm): + if self.vlm is None: + self.vlm = self.llm logging.info(f"[LLM] {self.current_llm} model loaded successfully.") - NGROK_TOKEN = os.environ.get("NGROK_TOKEN", "") if NGROK_TOKEN: ngrok.set_auth_token(NGROK_TOKEN) diff --git a/cuda.Dockerfile b/cuda.Dockerfile index bc33a5d..1caef07 100644 --- a/cuda.Dockerfile +++ b/cuda.Dockerfile @@ -15,7 +15,7 @@ COPY cuda-requirements.txt . RUN python3 -m pip install --upgrade pip cmake scikit-build setuptools wheel --no-cache-dir && \ CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python==0.2.55 --no-cache-dir && \ pip install --no-cache-dir -r cuda-requirements.txt -RUN git clone https://github.com/deepseek-ai/DeepSeek-VL deepseek && \ +RUN git clone https://github.com/Josh-XT/DeepSeek-VL deepseek && \ cd deepseek && \ pip install --no-cache-dir -e . && \ cd .. diff --git a/tests.ipynb b/tests.ipynb index 5f734dd..a23046c 100644 --- a/tests.ipynb +++ b/tests.ipynb @@ -23,7 +23,7 @@ }, { "cell_type": "code", - "execution_count": 12, + "execution_count": 24, "metadata": {}, "outputs": [], "source": [ @@ -112,7 +112,7 @@ }, { "cell_type": "code", - "execution_count": 13, + "execution_count": 25, "metadata": {}, "outputs": [ { @@ -147,7 +147,7 @@ }, { "cell_type": "code", - "execution_count": 14, + "execution_count": 26, "metadata": {}, "outputs": [ { @@ -174,7 +174,7 @@ }, { "cell_type": "code", - "execution_count": 15, + "execution_count": 27, "metadata": {}, "outputs": [ { @@ -205,7 +205,7 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": 28, "metadata": {}, "outputs": [ { @@ -255,7 +255,7 @@ }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 29, "metadata": {}, "outputs": [ { @@ -312,7 +312,7 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": 30, "metadata": {}, "outputs": [ { @@ -353,7 +353,7 @@ }, { "cell_type": "code", - "execution_count": 19, + "execution_count": 31, "metadata": {}, "outputs": [ { @@ -361,7 +361,7 @@ "text/html": [ "\n", " \n", " " @@ -370,7 +370,7 @@ "" ] }, - "execution_count": 19, + "execution_count": 31, "metadata": {}, "output_type": "execute_result" } @@ -407,7 +407,7 @@ }, { "cell_type": "code", - "execution_count": 20, + "execution_count": 32, "metadata": {}, "outputs": [ { @@ -434,7 +434,7 @@ }, { "cell_type": "code", - "execution_count": 21, + "execution_count": 33, "metadata": {}, "outputs": [ { @@ -469,7 +469,7 @@ }, { "cell_type": "code", - "execution_count": 22, + "execution_count": 34, "metadata": {}, "outputs": [ { @@ -477,7 +477,7 @@ "text/html": [ "\n", " \n", " "